AnalysisofSeawater A Guide for the Analytical and ...dl.mozh.org/up/Analysis_of_Seawate.pdf · T.R....

Analysis of SeawaterA Guide for the Analytical

and Environmental Chemist

T.R. Crompton

Analysisof SeawaterA Guide for the Analyticaland Environmental Chemist

123

With 48 Figures and 45 Tables

T.R. CromptonHill Cottage (Bwthyn Yr Allt)

Anglesey, Gwynedd

United Kingdom

Library of Congress Control Number: 2005929412

ISBN-10 3-540-26762-X Springer Berlin Heidelberg New York

DOI 10.1007/b95901

This work is subject to copyright. All rights reserved, whether the whole or part of the material is concerned,specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproductionon microfilm or in any other way, and storage in data banks. Duplication of this publication or parts thereofis permitted only under the provisions of the German Copyright Law of September 9, 1965, in its currentversion, and permission for use must always be obtained from Springer. Violations are liable for prosecutionunder the German Copyright Law.

Springer is a part of Springer Science+Business Media

© Springer-Verlag Berlin Heidelberg 2006Printed in Germany

The use of general descriptive names, registered names, trademarks, etc. in this publication does not imply,even in the absence of a specific statement, that such names are exempt from the relevant protective lawsand regulations and therefore free for general use.

Product liability: The publishers cannot guarantee the accuracy of any information about dosage andapplication contained in this book. In every individual case the user must check such information byconsulting the relevant literature.

Cover design: design & production GmbH, HeidelbergTypesetting and production: LE-TEX Jelonek, Schmidt & Vöckler GbR, Leipzig, GermanyPrinted on acid-free paper 2/3141/YL - 5 4 3 2 1 0

ISBN-13 978-3-540-26762-1 Springer Berlin Heidelberg New York

springer.com

Preface

This book covers all aspects of the analysis of seawater using both classical andthe most advanced recently introduced physical techniques.

Until fairly recently, the analysis of seawater was limited to a number ofmajor constituents such as chloride and alkalinity.

It was generally agreed that any determinations of trace metals carried outon seawater prior to about 1975 are questionable, principally due to the adverseeffects of contamination during sampling, which were then little understoodand lead to artificially high results. It is only in the past few years that meth-ods of adequate sensitivity have become available for true ultra-trace metaldeterminations in water.

Similar comments apply in the case of organics in seawater, because it hasnow become possible to resolve the complex mixtures of organics in seawaterand achieve the required very low detection limits. Only since the advent ofsample preconcentration and mass spectrometry coupled with gas chromatog-raphy and high-performance liquid chromatography, and possibly derivatisa-tion of the original sample constituents to convert them into a form suitablefor chromatography, has this become possible.

Fortunately, our interest in micro-constituents in the seawater both fromthe environmental and the nutrient balance points of view has coincided withthe availability of advanced instrumentation capable of meeting the analyticalneeds.

Chapter 1 discusses a very important aspect of seawater analysis, namelysampling. If the sample is not taken correctly, the final result is invalidated,no matter how sophisticated the final analytical procedure. Recent importantwork on sampling is discussed in detail.

Chapters 2 and 3 discuss the determination of anions. Direct application ofmanyof the classical procedures for anions fail for seawaterowing to interferingeffects of the simple matrix. Suitable modifications are discussed that areamenable to seawater. Dissolved gases in seawater are of interest in certaincontexts and their determination is discussed in Chap. 4.

Chapters 5 and 6 discuss the application of new techniques such as atomicabsorption spectrometry with and without graphite furnace and Zeeman back-ground correction, inductively coupled plasma mass spectrometry, X-ray fluo-

VI Preface

rescence spectrometry, neutron activation analysis, voltammetric techniques,and others. In the first part of the chapter elements are discussed singly in al-phabetical order, then as the groups of elements, because the newer techniquesoften cover ranges of elements. Finally, there is a section on metal preconcen-tration techniques. By concentrating all the metals present in a large volumeof sample into a few millilitres, dramatic improvements in detection limitedcan be achieved, and it is this that enables the techniques to be applicableat the low basal concentrations at which many metals exist in seawater. In-creasingly, owing to fallout and the introduction of nuclear power stations andprocessing plants, it is necessary to monitor the levels of radioactive elementsin the seawater, and this is discussed in Chap. 7. Chapters 8 and 9 cover thedetermination of a wide range of organics in seawater, whilst Chap. 10 coversorganometallic compounds. An increasingly long list of organometallics canbe detected in seawater, many of these being produced by biologically inducedmetal methylation processes occurring in sediment and fish tissues.

Finally, in Chap. 11, is discussed the present state of knowledge on the de-termination of various oxygen demand parameters and non-metallic elementsin seawater. Amongst others these include total, dissolved, and volatile organiccarbon, and total inorganic carbon, as well as recent work on the older oxygendemand parameters such as chemical oxygen demand and biochemical oxygendemand. The confusion that formerly existed regarding these methodologiesin now being resolved to the point that meaningful measurements can now bereported. The determination of other non-metallic elements is also discussedin this chapter.

Whilst the book will be of obvious interest to anyone concerned with sea-water environmental protection, it is believed that it will also be of interest toother groups of workers, including River Authorities who have to implementlegal requirements regarding seawater pollution, oceanographers, fisheriesexperts and politicians who create and implement environmental policies, andthe news media, who are responsible for making the general public aware ofenvironmental matters.

The book will also be of interest to practising analysts and, not least, tothe scientists and environmentalists of the future who are currently passingthrough the university system and on whom, more than even previously, willrest the responsibility of ensuring that our oceans are protected in the future.

Anglesey, UK, 2006 T.R. Crompton

Contents

1 Sampling and Storage . . . . . . . . . . . . . . . . . . . . . . . 11.1 Sampling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.2 Sampling Devices . . . . . . . . . . . . . . . . . . . . . . . . 31.2.1 Intercomparison of Seawater Sampling for Trace Metals . . . 71.2.2 Intercomparison of Sampling Devices

and Analytical Techniques Using Seawater from a CEPEX(Controlled, Ecosystem Pollution Experiment) Enclosure . . 12

1.3 Sample Preservation and Storage . . . . . . . . . . . . . . . . 171.3.1 Losses of Silver, Arsenic, Cadmium, Selenium, and Zinc

from Seawater by Sorptionon Various Container Surfaces [54] . . . . . . . . . . . . . . . 19

1.3.2 Losses of Phthalic Acid Esters and Polychlorinated Biphenylsfrom Seawater Samples During Storage . . . . . . . . . . . . 26

1.4 Sample Contamination During Analysis . . . . . . . . . . . . 27References . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

2 Determination of Anions . . . . . . . . . . . . . . . . . . . . . 392.1 Acetate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 392.1.1 Ion Chromatography . . . . . . . . . . . . . . . . . . . . . . 392.2 Acrylate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 392.2.1 Ion Chromatography . . . . . . . . . . . . . . . . . . . . . . 392.3 Alkalinity . . . . . . . . . . . . . . . . . . . . . . . . . . . . 392.3.1 Titration Method . . . . . . . . . . . . . . . . . . . . . . . . 392.3.2 Spectrophotometric Methods . . . . . . . . . . . . . . . . . . 402.4 Arsenate/Arsenite . . . . . . . . . . . . . . . . . . . . . . . . 412.4.1 Spectrophotometric Method . . . . . . . . . . . . . . . . . . 412.5 Benzoate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 412.5.1 Ion Chromatography . . . . . . . . . . . . . . . . . . . . . . 412.6 Butyrate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 412.6.1 Ion Chromatography . . . . . . . . . . . . . . . . . . . . . . 412.7 Borate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 422.7.1 Spectrophotometric Method . . . . . . . . . . . . . . . . . . 422.8 Bromate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 422.8.1 Spectrophotometric Titration and Differential Pulse

Polarography . . . . . . . . . . . . . . . . . . . . . . . . . . . 42

VIII Contents

2.9 Bromide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 452.9.1 Titration Method . . . . . . . . . . . . . . . . . . . . . . . . 452.9.2 X-ray Emission Spectrometry . . . . . . . . . . . . . . . . . . 452.9.3 Segmented Flow Analysis . . . . . . . . . . . . . . . . . . . . 462.9.4 Solid State Membrane Electrodes . . . . . . . . . . . . . . . . 462.9.5 X-ray Fluorescence Spectroscopy . . . . . . . . . . . . . . . . 462.9.6 Isotachoelectrophoresis . . . . . . . . . . . . . . . . . . . . . 462.10 Chloride . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 472.10.1 Titration Method . . . . . . . . . . . . . . . . . . . . . . . . 472.10.2 Ion Selective Electrodes . . . . . . . . . . . . . . . . . . . . . 472.10.3 Chronopotentiometry . . . . . . . . . . . . . . . . . . . . . . 482.10.4 Miscellaneous . . . . . . . . . . . . . . . . . . . . . . . . . . 482.11 Chromate and Dichromate . . . . . . . . . . . . . . . . . . . 482.11.1 Atomic Absorption Spectrometry . . . . . . . . . . . . . . . 482.11.2 Organic Forms of Chromium . . . . . . . . . . . . . . . . . . 492.12 Fluoride . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 532.12.1 Spectrophotometric Method . . . . . . . . . . . . . . . . . . 532.12.2 Ion Selective Electrodes . . . . . . . . . . . . . . . . . . . . . 532.12.3 Photoactivation Analysis . . . . . . . . . . . . . . . . . . . . 562.12.4 Atomic Absorption Spectrometry . . . . . . . . . . . . . . . 562.13 Formate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 572.13.1 High Performance Liquid Chromatography (HPLC) . . . . . 572.14 Hypochlorite . . . . . . . . . . . . . . . . . . . . . . . . . . . 582.14.1 Spectrophotometric Method . . . . . . . . . . . . . . . . . . 582.15 Iodate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 582.15.1 Spectrophotometric Method . . . . . . . . . . . . . . . . . . 582.16 Iodide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 622.16.1 Titration Method . . . . . . . . . . . . . . . . . . . . . . . . 622.16.2 Spectrophotometric Method . . . . . . . . . . . . . . . . . . 632.16.3 Cathodic Stripping Voltammetry . . . . . . . . . . . . . . . . 632.16.4 Ion Chromatograpy . . . . . . . . . . . . . . . . . . . . . . . 642.16.5 Miscellaneous . . . . . . . . . . . . . . . . . . . . . . . . . . 642.17 Molybdate . . . . . . . . . . . . . . . . . . . . . . . . . . . . 652.17.1 Atomic Absorption Spectrometry . . . . . . . . . . . . . . . 652.18 Nitrate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 652.18.1 Spectrophotometric Methods . . . . . . . . . . . . . . . . . . 652.18.2 Ultraviolet Spectroscopy . . . . . . . . . . . . . . . . . . . . 662.18.3 Chemiluminescence Method . . . . . . . . . . . . . . . . . . 682.18.4 Flow Injection Analysis . . . . . . . . . . . . . . . . . . . . . 682.18.5 Continuous Flow Analysis . . . . . . . . . . . . . . . . . . . . 692.18.6 Cathodic Stripping Voltammetry . . . . . . . . . . . . . . . . 692.18.7 Ion Chromatography . . . . . . . . . . . . . . . . . . . . . . 692.18.8 Bacteriological Method . . . . . . . . . . . . . . . . . . . . . 692.18.9 Miscellaneous . . . . . . . . . . . . . . . . . . . . . . . . . . 71

Contents IX

2.19 Nitrite . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 712.19.1 Spectrophotometric Methods . . . . . . . . . . . . . . . . . . 712.19.2 Flow Injection Analysis . . . . . . . . . . . . . . . . . . . . . 722.19.3 Isotope Dilution Gas Chromatography . . . . . . . . . . . . . 722.19.4 Cathodic Stripping Voltammetry . . . . . . . . . . . . . . . . 722.20 Nitrate and Nitrite . . . . . . . . . . . . . . . . . . . . . . . . 732.20.1 Spectrophotometric Method . . . . . . . . . . . . . . . . . . 732.20.2 Flow Injection Analysis . . . . . . . . . . . . . . . . . . . . . 732.20.3 Continuous Flow Analysis . . . . . . . . . . . . . . . . . . . . 752.20.4 Reverse Phase Ion Interaction Liquid Chromatography . . . . 752.20.5 Miscellaneous . . . . . . . . . . . . . . . . . . . . . . . . . . 752.21 Perrhenate . . . . . . . . . . . . . . . . . . . . . . . . . . . . 762.22 Phosphate . . . . . . . . . . . . . . . . . . . . . . . . . . . . 762.22.1 Reverse Flow Injection Analysis . . . . . . . . . . . . . . . . 762.22.2 Spectrophotometric Method . . . . . . . . . . . . . . . . . . 772.22.3 Ion Chromatography . . . . . . . . . . . . . . . . . . . . . . 822.23 Propionate . . . . . . . . . . . . . . . . . . . . . . . . . . . . 822.23.1 Ion Chromatography . . . . . . . . . . . . . . . . . . . . . . 822.24 Pyruvate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 822.24.1 Ion Chromatography . . . . . . . . . . . . . . . . . . . . . . 822.25 Selenate/Selenite . . . . . . . . . . . . . . . . . . . . . . . . . 822.25.1 Fluorometric Method . . . . . . . . . . . . . . . . . . . . . . 822.26 Silicate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 832.26.1 Spectrophotometric Methods . . . . . . . . . . . . . . . . . . 832.26.2 Flow Injection Analysis . . . . . . . . . . . . . . . . . . . . . 842.26.3 Ion Exclusion Chromatography . . . . . . . . . . . . . . . . . 842.27 Sulfide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 852.27.1 Gas Chromatography . . . . . . . . . . . . . . . . . . . . . . 852.27.2 Capillary Isotachoelectrophoresis . . . . . . . . . . . . . . . 852.28 Sulfate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 862.28.1 Titration Method . . . . . . . . . . . . . . . . . . . . . . . . 862.28.2 Inductively Coupled Plasma Atomic Emission Spectrometry . 862.28.3 Polarography . . . . . . . . . . . . . . . . . . . . . . . . . . . 872.28.4 Ion Chromatography . . . . . . . . . . . . . . . . . . . . . . 882.29 Valerate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 882.29.1 Ion Chromatography . . . . . . . . . . . . . . . . . . . . . . 882.30 Multianion Analysis . . . . . . . . . . . . . . . . . . . . . . . 882.30.1 Spectrophotometric Methods, Phosphate, Arsenate, Arsenite,

and Sulfide . . . . . . . . . . . . . . . . . . . . . . . . . . . . 882.30.2 Electrostatic Ion Chromatography, Bromide, Nitrate,

and Iodide . . . . . . . . . . . . . . . . . . . . . . . . . . . . 892.30.3 Miscellaneous . . . . . . . . . . . . . . . . . . . . . . . . . . 902.31 pH . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 902.32 Suspended Solids . . . . . . . . . . . . . . . . . . . . . . . . 91

X Contents

2.33 Anion Preconcentration . . . . . . . . . . . . . . . . . . . . . 92References . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92

3 Anions in Estuary and Coastal Waters . . . . . . . . . . . . . . . 993.1 Nitrate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 993.1.1 Ultraviolet Spectroscopy . . . . . . . . . . . . . . . . . . . . 993.2 Nitrate and Nitrite . . . . . . . . . . . . . . . . . . . . . . . . 993.2.1 Autoanalyser Method . . . . . . . . . . . . . . . . . . . . . . 993.3 Phosphate . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1003.3.1 Spectrophotometric Method . . . . . . . . . . . . . . . . . . 1003.4 Selenate and Selenite . . . . . . . . . . . . . . . . . . . . . . 1003.4.1 Spectrofluorometric Method . . . . . . . . . . . . . . . . . . 1003.4.2 Atomic Absorption Spectrometry . . . . . . . . . . . . . . . 1013.5 Sulfate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1013.5.1 Spectrophotometric Method . . . . . . . . . . . . . . . . . . 1013.6 Multianion Analysis . . . . . . . . . . . . . . . . . . . . . . . 1023.6.1 Spectrophotometric Method, Sulfate, Phosphate, Nitrate,

and Sulfide . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102References . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102

4 Dissolved Gases . . . . . . . . . . . . . . . . . . . . . . . . . . 1034.1 Free Chlorine . . . . . . . . . . . . . . . . . . . . . . . . . . 1034.1.1 Amperometric Titration Procedures . . . . . . . . . . . . . . 1034.2 Ozone . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1084.3 Nitric Oxide . . . . . . . . . . . . . . . . . . . . . . . . . . . 1084.4 Hydrogen Sulfide . . . . . . . . . . . . . . . . . . . . . . . . 1084.5 Carbon Dioxide . . . . . . . . . . . . . . . . . . . . . . . . . 108

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109

5 Cations in Seawater . . . . . . . . . . . . . . . . . . . . . . . . 1115.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 1115.2 Actinium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1125.3 Aluminium . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1125.3.1 Spectrophotometric Methods . . . . . . . . . . . . . . . . . . 1125.3.2 Spectrofluorometric Methods . . . . . . . . . . . . . . . . . . 1135.3.3 Atomic Absorption Spectrometry . . . . . . . . . . . . . . . 1145.3.4 Anodic Stripping Voltammetry . . . . . . . . . . . . . . . . . 1145.3.5 Gas Chromatography . . . . . . . . . . . . . . . . . . . . . . 1145.4 Ammonium . . . . . . . . . . . . . . . . . . . . . . . . . . . 1155.4.1 Spectrophotometric Methods . . . . . . . . . . . . . . . . . . 1155.4.2 Flow Injection Analysis . . . . . . . . . . . . . . . . . . . . . 1185.4.3 Ion-Selective Electrodes . . . . . . . . . . . . . . . . . . . . . 1185.4.4 High-Performance Liquid Chromatography . . . . . . . . . . 1185.5 Antimony . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119

Contents XI

5.5.1 Atomic Absorption Spectrometry . . . . . . . . . . . . . . . 1195.5.2 Hydride Generation Atomic Absorption Spectrometry . . . . 1195.6 Arsenic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1205.6.1 Spectrophotometric Methods . . . . . . . . . . . . . . . . . . 1205.6.2 Atomic Absorption Spectrometry . . . . . . . . . . . . . . . 1215.6.3 Neutron Activation Analysis . . . . . . . . . . . . . . . . . . 1225.6.4 Inductively Coupled Plasma Mass Spectrometry . . . . . . . 1235.6.5 Anodic Stripping Voltammetry . . . . . . . . . . . . . . . . . 1235.6.6 X-ray Fluorescence Spectroscopy . . . . . . . . . . . . . . . . 1245.7 Barium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1245.7.1 Atomic Absorption Spectrometry . . . . . . . . . . . . . . . 1245.8 Beryllium . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1255.8.1 Graphite Furnace Atomic Absorption Spectrometry . . . . . 1255.8.2 Miscellaneous . . . . . . . . . . . . . . . . . . . . . . . . . . 1255.9 Bismuth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1265.9.1 Atomic Absorption Spectrometry . . . . . . . . . . . . . . . 1265.10 Boron . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1275.10.1 Spectrophotometric Methods . . . . . . . . . . . . . . . . . . 1275.10.2 Atomic Absorption Spectrometry . . . . . . . . . . . . . . . 1285.10.3 Coulometry . . . . . . . . . . . . . . . . . . . . . . . . . . . 1285.11 Cadmium . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1295.11.1 Atomic Absorption Spectrometry . . . . . . . . . . . . . . . 1295.11.2 Anodic Stripping Voltammetry . . . . . . . . . . . . . . . . . 1345.12 Caesium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1355.12.1 Atomic Absorption Spectrometry . . . . . . . . . . . . . . . 1355.13 Cerium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1365.14 Calcium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1365.14.1 Titration Methods . . . . . . . . . . . . . . . . . . . . . . . . 1365.14.2 Atomic Absorption Spectrometry . . . . . . . . . . . . . . . 1385.14.3 Flame Photometry . . . . . . . . . . . . . . . . . . . . . . . . 1385.14.4 Calcium-Selective Electrodes . . . . . . . . . . . . . . . . . . 1385.14.5 Inductively Coupled Plasma Atomic Emission Spectrometry . 1395.15 Chromium . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1395.15.1 Total Chromium . . . . . . . . . . . . . . . . . . . . . . . . . 1395.15.2 Chromium (III) . . . . . . . . . . . . . . . . . . . . . . . . . 1425.15.3 Chromium (III) and (VI) . . . . . . . . . . . . . . . . . . . . 1435.15.4 Chromium (III) and Total Chromium. Gas Chromatography . 1455.15.5 Organic Forms of Chromium . . . . . . . . . . . . . . . . . . 1455.16 Cobalt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1485.16.1 Spectrophotometric Methods . . . . . . . . . . . . . . . . . . 1485.16.2 Atomic Absorption Spectrometry . . . . . . . . . . . . . . . 1495.16.3 Flow Injection Analysis . . . . . . . . . . . . . . . . . . . . . 1505.16.4 Atomic Fluorescence Spectrometry . . . . . . . . . . . . . . . 1505.16.5 Spectrofluorometry . . . . . . . . . . . . . . . . . . . . . . . 150

XII Contents

5.16.6 Chemical Luminescence Analysis . . . . . . . . . . . . . . . . 1505.16.7 Cathodic Stripping Voltammetry . . . . . . . . . . . . . . . . 1515.16.8 Polarography . . . . . . . . . . . . . . . . . . . . . . . . . . . 1515.17 Copper . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1525.17.1 Titration Procedures . . . . . . . . . . . . . . . . . . . . . . 1535.17.2 Atomic Absorption Spectrometry . . . . . . . . . . . . . . . 1545.17.3 Spectrophotometric Method and Spectrofluorometric Method 1555.17.4 Ion-Selective Electrodes . . . . . . . . . . . . . . . . . . . . . 1555.17.5 Electroanalytical Methods . . . . . . . . . . . . . . . . . . . 1555.17.6 Isotope Dilution Methods . . . . . . . . . . . . . . . . . . . . 1575.17.7 Electron Spin Resonance Spectrometry . . . . . . . . . . . . 1575.17.8 Miscellaneous Methods . . . . . . . . . . . . . . . . . . . . . 1575.17.9 Copper Speciation . . . . . . . . . . . . . . . . . . . . . . . . 1575.18 Dysprosium . . . . . . . . . . . . . . . . . . . . . . . . . . . 1635.19 Erbium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1635.20 Europium . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1635.21 Gadolinium . . . . . . . . . . . . . . . . . . . . . . . . . . . 1635.22 Gallium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1635.23 Germanium . . . . . . . . . . . . . . . . . . . . . . . . . . . 1635.23.1 Hydride Generation Furnace Atomic Absorption

Spectrometry . . . . . . . . . . . . . . . . . . . . . . . . . . 1635.24 Gold . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1645.24.1 Inductively Coupled Plasma Mass Spectrometry . . . . . . . 1645.24.2 Photometry . . . . . . . . . . . . . . . . . . . . . . . . . . . 1645.25 Holmium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1645.26 Indium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1645.26.1 Neutron Activation Analysis . . . . . . . . . . . . . . . . . . 1645.27 Iridium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1655.28 Iron . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1655.28.1 Spectrophotometric Methods . . . . . . . . . . . . . . . . . . 1655.28.2 Atomic Absorption Spectrometry . . . . . . . . . . . . . . . 1665.28.3 Chemiluminescence . . . . . . . . . . . . . . . . . . . . . . . 1665.28.4 Voltammetry . . . . . . . . . . . . . . . . . . . . . . . . . . . 1675.28.5 Radioisotope Dilution . . . . . . . . . . . . . . . . . . . . . . 1675.29 Lanthanum . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1675.30 Lead . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1685.30.1 Atomic Fluorescence Spectroscopy . . . . . . . . . . . . . . . 1685.30.2 Flow Injection Analysis . . . . . . . . . . . . . . . . . . . . . 1685.30.3 Atomic Absorption Spectrometry . . . . . . . . . . . . . . . 1685.30.4 Anodic Stripping Voltammetry . . . . . . . . . . . . . . . . . 1725.30.5 Mass Spectrometry . . . . . . . . . . . . . . . . . . . . . . . 1745.30.6 Miscellaneous . . . . . . . . . . . . . . . . . . . . . . . . . . 1745.31 Lithium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1745.31.1 Atomic Absorption Spectrometry . . . . . . . . . . . . . . . 174

Contents XIII

5.31.2 Gel Permeation Chromatography . . . . . . . . . . . . . . . . 1745.31.3 Neutron Activation Analysis . . . . . . . . . . . . . . . . . . 1745.32 Lutetium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1755.33 Magnesium . . . . . . . . . . . . . . . . . . . . . . . . . . . 1755.33.1 Gravimetric Method . . . . . . . . . . . . . . . . . . . . . . . 1755.33.2 Atomic Absorption Spectrometry . . . . . . . . . . . . . . . 1755.34 Manganese . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1755.34.1 Spectrophotometric Methods . . . . . . . . . . . . . . . . . . 1765.34.2 Spectrofluorometric Method . . . . . . . . . . . . . . . . . . 1775.34.3 Atomic Absorption Spectrometry . . . . . . . . . . . . . . . 1775.34.4 Polarography . . . . . . . . . . . . . . . . . . . . . . . . . . . 1805.34.5 Neutron Activation Analysis . . . . . . . . . . . . . . . . . . 1805.35 Mercury . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1805.35.1 Atomic Absorption Spectrometry . . . . . . . . . . . . . . . 1805.35.2 Inductively Coupled Plasma Mass Spectrometry . . . . . . . 1845.35.3 Inductively Coupled Plasma Atomic Emission Spectrometry . 1845.35.4 Atomic Emission Spectrometry . . . . . . . . . . . . . . . . . 1845.35.5 Colloid Flotation . . . . . . . . . . . . . . . . . . . . . . . . . 1845.35.6 Miscellaneous . . . . . . . . . . . . . . . . . . . . . . . . . . 1865.36 Molybdenum . . . . . . . . . . . . . . . . . . . . . . . . . . . 1865.36.1 Spectrophotometric Methods . . . . . . . . . . . . . . . . . . 1865.36.2 Atomic Absorption Spectrometry . . . . . . . . . . . . . . . 1875.36.3 Inductively Coupled Plasma Mass Spectrometry . . . . . . . 1885.36.4 Electrochemical Methods . . . . . . . . . . . . . . . . . . . . 1885.36.5 X-ray Fluorescence Spectrometry . . . . . . . . . . . . . . . . 1895.36.6 Miscellaneous . . . . . . . . . . . . . . . . . . . . . . . . . . 1895.37 Neodymium . . . . . . . . . . . . . . . . . . . . . . . . . . . 1895.38 Neptunium . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1905.39 Nickel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1905.39.1 Spectrophotometric Method . . . . . . . . . . . . . . . . . . 1905.39.2 Atomic Absorption Spectrometry . . . . . . . . . . . . . . . 1905.39.3 Cathodic Stripping Voltammetry . . . . . . . . . . . . . . . . 1915.39.4 Liquid Scintillation Counting . . . . . . . . . . . . . . . . . . 1925.40 Osmium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1925.40.1 Resonance Ionisation Mass Spectrometry . . . . . . . . . . . 1925.41 Palladium . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1925.42 Platinum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1925.42.1 Cathodic Stripping Voltammetry . . . . . . . . . . . . . . . . 1925.43 Plutonium . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1925.44 Polonium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1935.45 Potassium . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1935.45.1 Titration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1935.45.2 Polarography . . . . . . . . . . . . . . . . . . . . . . . . . . . 1935.45.3 Ion-Selective Electrodes . . . . . . . . . . . . . . . . . . . . . 194

XIV Contents

5.46 Praseodymium . . . . . . . . . . . . . . . . . . . . . . . . . . 1945.47 Promethium . . . . . . . . . . . . . . . . . . . . . . . . . . . 1945.48 Radium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1945.49 Rare Earths . . . . . . . . . . . . . . . . . . . . . . . . . . . 1945.49.1 Cerium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1945.49.2 Praseodymium . . . . . . . . . . . . . . . . . . . . . . . . . . 1955.49.3 Neodymium . . . . . . . . . . . . . . . . . . . . . . . . . . . 1955.49.4 Promethium . . . . . . . . . . . . . . . . . . . . . . . . . . . 1955.49.5 Samarium . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1955.49.6 Europium . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1965.49.7 Gadolinium . . . . . . . . . . . . . . . . . . . . . . . . . . . 1965.49.8 Terbium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1965.49.9 Dysprosium . . . . . . . . . . . . . . . . . . . . . . . . . . . 1965.49.10 Holmium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1965.49.11 Erbium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1965.49.12 Thulium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1965.49.13 Ytterbium . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1965.49.14 Lutetium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1965.49.15 Analysis of Rare Earth Mixtures . . . . . . . . . . . . . . . . 1975.50 Rhenium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1995.50.1 Graphite Furnace Atomic Absorption Spectrometry . . . . . 1995.50.2 Neutron Activation Analysis . . . . . . . . . . . . . . . . . . 2005.51 Rubidium . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2005.51.1 Atomic Absorption Spectrometry . . . . . . . . . . . . . . . 2005.51.2 Spectrometry . . . . . . . . . . . . . . . . . . . . . . . . . . 2015.51.3 Mass Spectrometry . . . . . . . . . . . . . . . . . . . . . . . 2015.51.4 X-ray Fluorescence Spectroscopy . . . . . . . . . . . . . . . . 2015.52 Ruthenium . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2015.53 Samarium . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2015.54 Scandium . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2015.55 Selenium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2015.55.1 Spectrophotometry . . . . . . . . . . . . . . . . . . . . . . . 2025.55.2 Atomic Absorption Spectrometry . . . . . . . . . . . . . . . 2025.55.3 Hydride Generation Atomic Absorption Spectrometry . . . . 2025.55.4 Cathodic Stripping Voltammetry . . . . . . . . . . . . . . . . 2025.55.5 Gas Chromatography . . . . . . . . . . . . . . . . . . . . . . 2035.55.6 Neutron Activation Analysis . . . . . . . . . . . . . . . . . . 2035.56 Silver . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2035.56.1 Atomic Absorption Spectrometry . . . . . . . . . . . . . . . 2035.56.2 Neutron Activation Analysis . . . . . . . . . . . . . . . . . . 2045.57 Sodium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2045.57.1 Amperometry . . . . . . . . . . . . . . . . . . . . . . . . . . 2045.57.2 Polarimetry . . . . . . . . . . . . . . . . . . . . . . . . . . . 2045.58 Strontium . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205

Contents XV

5.58.1 Atomic Absorption Spectrometry . . . . . . . . . . . . . . . 2055.59 Technetium . . . . . . . . . . . . . . . . . . . . . . . . . . . 2055.60 Tellurium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2055.60.1 Atomic Absorption Spectrometry . . . . . . . . . . . . . . . 2055.61 Terbium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2065.62 Thallium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2065.63 Thorium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2065.63.1 Thermal Ion Mass Spectrometry . . . . . . . . . . . . . . . . 2065.63.2 Neutron Activation Analysis . . . . . . . . . . . . . . . . . . 2065.64 Thulium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2075.65 Tin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2075.65.1 Spectrophotometric Method . . . . . . . . . . . . . . . . . . 2075.65.2 Atomic Absorption Spectrometry . . . . . . . . . . . . . . . 2075.65.3 Gas Chromatography . . . . . . . . . . . . . . . . . . . . . . 2075.65.4 High-Performance Liquid Chromatography . . . . . . . . . . 2095.65.5 Anodic Stripping Voltammetry . . . . . . . . . . . . . . . . . 2105.65.6 Miscellaneous . . . . . . . . . . . . . . . . . . . . . . . . . . 2115.66 Titanium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2115.66.1 Spectrophotometric Method . . . . . . . . . . . . . . . . . . 2115.67 Tungsten . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2115.68 Uranium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2115.68.1 Spectrophotometric Method . . . . . . . . . . . . . . . . . . 2115.68.2 Cathodic Stripping Voltammetry . . . . . . . . . . . . . . . . 2115.68.3 Polarography . . . . . . . . . . . . . . . . . . . . . . . . . . . 2125.68.4 Miscellaneous . . . . . . . . . . . . . . . . . . . . . . . . . . 2125.69 Vanadium . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2135.69.1 Spectrophotometric Method . . . . . . . . . . . . . . . . . . 2135.69.2 Atomic Absorption Spectrometry . . . . . . . . . . . . . . . 2135.69.3 Inductively Coupled Plasma Mass Spectrometry . . . . . . . 2145.69.4 Cathodic Stripping Voltammetry . . . . . . . . . . . . . . . . 2145.69.5 Neutron Activation Analysis . . . . . . . . . . . . . . . . . . 2145.70 Ytterbium . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2155.71 Yttrium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2155.72 Zinc . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2155.72.1 Spectrofluorometric Method . . . . . . . . . . . . . . . . . . 2165.72.2 Atomic Absorption Spectrometry . . . . . . . . . . . . . . . 2165.72.3 Flow Injection Analysis . . . . . . . . . . . . . . . . . . . . . 2175.72.4 Stripping Voltammetry . . . . . . . . . . . . . . . . . . . . . 2175.72.5 Miscellaneous . . . . . . . . . . . . . . . . . . . . . . . . . . 2185.73 Zirconium . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2185.74 Multication Analysis . . . . . . . . . . . . . . . . . . . . . . . 2185.74.1 Titration Procedures . . . . . . . . . . . . . . . . . . . . . . 2185.74.2 Spectrophotometric Procedure . . . . . . . . . . . . . . . . . 2195.74.3 Molecular Photoluminescence Spectrometry . . . . . . . . . 219

XVI Contents

5.74.4 Flame Atomic Absorption Spectrometry . . . . . . . . . . . . 2205.74.5 Graphite Furnace Atomic Absorption Spectrometry . . . . . 2235.74.6 Zeeman Graphite Furnace Atomic Absorption Spectrometry . 2315.74.7 Hydride Generation Atomic Absorption Spectrometry . . . . 2335.74.8 Inductively Coupled Plasma Atomic Emission Spectrometry . 2405.74.9 Inductively Coupled Plasma Mass Spectrometry . . . . . . . 2445.74.10 Plasma Emission Spectrometry . . . . . . . . . . . . . . . . . 2485.74.11 Anodic Stripping Voltammetry . . . . . . . . . . . . . . . . . 2485.74.12 Cathodic Stripping Voltammetry . . . . . . . . . . . . . . . . 2595.74.13 Chronopotentiometry . . . . . . . . . . . . . . . . . . . . . . 2605.74.14 X-ray Fluorescence Spectrometry . . . . . . . . . . . . . . . . 2615.74.15 Neutron Activation Analysis . . . . . . . . . . . . . . . . . . 2625.74.16 Isotope Dilution Mass Spectrometry . . . . . . . . . . . . . . 2685.74.17 High-Performance Liquid Chromatography . . . . . . . . . . 2715.74.18 Metal Speciation . . . . . . . . . . . . . . . . . . . . . . . . . 2715.74.19 Metal Preconcentration . . . . . . . . . . . . . . . . . . . . . 2855.74.20 Miscellaneous . . . . . . . . . . . . . . . . . . . . . . . . . . 288

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . 288

6 Cations in Estuary, Bay, and Coastal Waters . . . . . . . . . . . . 3136.1 Ammonium . . . . . . . . . . . . . . . . . . . . . . . . . . . 3136.2 Arsenic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3146.2.1 Hydride Generation Atomic Spectrometry . . . . . . . . . . . 3146.3 Barium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3146.3.1 Atomic Absorption Spectrometry . . . . . . . . . . . . . . . 3146.4 Cadmium . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3156.4.1 Atomic Absorption Spectrometry . . . . . . . . . . . . . . . 3156.5 Calcium and Magnesium . . . . . . . . . . . . . . . . . . . . 3166.6 Copper . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3166.6.1 Titration Procedure . . . . . . . . . . . . . . . . . . . . . . . 3166.6.2 Anodic Stripping Voltammetry . . . . . . . . . . . . . . . . . 3166.7 Mercury . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3176.7.1 Miscellaneous . . . . . . . . . . . . . . . . . . . . . . . . . . 3176.8 Manganese . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3186.8.1 Polarography . . . . . . . . . . . . . . . . . . . . . . . . . . . 3186.9 Selenium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3186.9.1 Hydride Generation Graphite Furnace

Atomic Absorption Spectrometry . . . . . . . . . . . . . . . 3186.10 Tin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3186.10.1 High-Performance Liquid Chromatography . . . . . . . . . . 3186.11 Multication Analysis . . . . . . . . . . . . . . . . . . . . . . . 3196.11.1 Heavy Metals, Isotope Dilution, Spark Source

Mass Spectrometry, and Inductively Coupled Plasma AtomicEmission Spectrometry . . . . . . . . . . . . . . . . . . . . . 319

Contents XVII

6.11.2 Anodic Stripping Voltammetry . . . . . . . . . . . . . . . . . 3226.11.3 Cathodic Stripping Voltammetry . . . . . . . . . . . . . . . . 3226.11.4 Emission Spectrometry . . . . . . . . . . . . . . . . . . . . . 3236.11.5 Hydride Generation Atomic Spectrometry . . . . . . . . . . . 3236.11.6 Inductively Coupled Plasma Mass Spectrometry . . . . . . . 3236.11.7 Preconcentration Techniques . . . . . . . . . . . . . . . . . . 3246.11.8 Speciation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 325

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . 325

7 Radioactive Elements . . . . . . . . . . . . . . . . . . . . . . . 3297.1 Naturally Occurring Cations . . . . . . . . . . . . . . . . . . 3297.1.1 Actinium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3297.1.2 Polonium and Lead . . . . . . . . . . . . . . . . . . . . . . . 3297.1.3 Radium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3317.1.3.1 Radium, Barium, and Radon . . . . . . . . . . . . . . . . . . 3317.1.3.2 Radium, Thorium, and Lead . . . . . . . . . . . . . . . . . . 3327.1.4 99Technetium . . . . . . . . . . . . . . . . . . . . . . . . . . 3337.1.5 Thorium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3337.1.6 Bromide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3357.1.7 Phosphate . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3357.2 Fallout Products and Nuclear Plant Emissions . . . . . . . . . 3367.2.1 Americium and Plutonium . . . . . . . . . . . . . . . . . . . 3367.2.2 137Caesium . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3367.2.3 60Cobalt . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3387.2.4 55Iron . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3387.2.5 54Manganese . . . . . . . . . . . . . . . . . . . . . . . . . . . 3387.2.6 237Neptunium . . . . . . . . . . . . . . . . . . . . . . . . . . 3397.2.7 Plutonium . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3397.2.8 106Ruthenium and Osmium . . . . . . . . . . . . . . . . . . . 3417.2.9 90Strontium . . . . . . . . . . . . . . . . . . . . . . . . . . . 3417.2.10 Uranium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3427.2.11 Miscellaneous . . . . . . . . . . . . . . . . . . . . . . . . . . 344

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . 344

8 Sample Preparation Prior to Analysis for Organics . . . . . . . . . 3498.1 Soluble Components of Seawater . . . . . . . . . . . . . . . . 3508.1.1 Reverse Osmosis . . . . . . . . . . . . . . . . . . . . . . . . . 3508.1.2 Freeze Drying . . . . . . . . . . . . . . . . . . . . . . . . . . 3508.1.3 Freezing-Out Methods . . . . . . . . . . . . . . . . . . . . . 3518.1.4 Froth Flotation . . . . . . . . . . . . . . . . . . . . . . . . . . 3518.1.5 Solvent Extraction . . . . . . . . . . . . . . . . . . . . . . . . 3518.1.6 Coprecipitation Techniques . . . . . . . . . . . . . . . . . . . 3538.1.7 Adsorption Techniques . . . . . . . . . . . . . . . . . . . . . 3548.2 Volatile Compounds of Seawater . . . . . . . . . . . . . . . . 355

XVIII Contents

8.2.1 Gas Stripping . . . . . . . . . . . . . . . . . . . . . . . . . . 3558.2.2 Headspace Analysis . . . . . . . . . . . . . . . . . . . . . . . 3578.2.3 Fractionation . . . . . . . . . . . . . . . . . . . . . . . . . . 3588.3 Chemical Pretreatment of Organics . . . . . . . . . . . . . . 361

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . 362

9 Organic Compounds . . . . . . . . . . . . . . . . . . . . . . . . 3659.1 Aliphatic Hydrocarbons . . . . . . . . . . . . . . . . . . . . . 3669.1.1 Spectrofluorometry . . . . . . . . . . . . . . . . . . . . . . . 3669.1.2 Dynamic Headspace Analysis . . . . . . . . . . . . . . . . . . 3669.1.3 Raman Spectroscopy . . . . . . . . . . . . . . . . . . . . . . 3689.1.4 Flow Calorimetry . . . . . . . . . . . . . . . . . . . . . . . . 3689.2 Aromatic Hydrocarbons . . . . . . . . . . . . . . . . . . . . . 3689.2.1 Spectrofluorometry . . . . . . . . . . . . . . . . . . . . . . . 3689.2.2 High-Performance Liquid Chromatography (HPLC) . . . . . 3699.3 Polyaromatic Hydrocarbons . . . . . . . . . . . . . . . . . . 3699.4 Oil Spills . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3709.4.1 Spectrofluorometry . . . . . . . . . . . . . . . . . . . . . . . 3709.4.2 Infrared Spectroscopy . . . . . . . . . . . . . . . . . . . . . . 3719.4.3 Gas Chromatography . . . . . . . . . . . . . . . . . . . . . . 3739.4.4 Gas Chromatography–Mass Spectrometry (GC–MS) . . . . . 3759.4.5 Miscellaneous . . . . . . . . . . . . . . . . . . . . . . . . . . 3779.5 Carboxylic Acids and Hydroxy Acids . . . . . . . . . . . . . . 3779.5.1 Spectrophotometric Method . . . . . . . . . . . . . . . . . . 3779.5.2 Gas Chromatography . . . . . . . . . . . . . . . . . . . . . . 3779.5.3 Liquid Chromatography . . . . . . . . . . . . . . . . . . . . . 3789.5.4 Atomic Absorption Spectrometry (AAS) . . . . . . . . . . . . 3799.5.5 Diffusion Method . . . . . . . . . . . . . . . . . . . . . . . . 3799.6 Ketones and Aldehydes . . . . . . . . . . . . . . . . . . . . . 3809.6.1 Spectrophotometric Method, Fluorometric

and Chemiluminescence Methods . . . . . . . . . . . . . . . 3809.6.2 Potential Sweep Voltammetry . . . . . . . . . . . . . . . . . . 3809.6.3 Gas Chromatography . . . . . . . . . . . . . . . . . . . . . . 3819.7 Phenols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3819.7.1 Spectrophotometric Methods . . . . . . . . . . . . . . . . . . 3819.7.2 Gas Chromatography–Mass Spectrometry (GC–MS) . . . . . 3829.8 Phthalate Esters . . . . . . . . . . . . . . . . . . . . . . . . . 3829.9 Carbohydrates . . . . . . . . . . . . . . . . . . . . . . . . . . 3829.9.1 Spectrophotometry . . . . . . . . . . . . . . . . . . . . . . . 3829.9.2 Enzymic Methods . . . . . . . . . . . . . . . . . . . . . . . . 3849.9.3 Liquid Chromatography . . . . . . . . . . . . . . . . . . . . 3849.9.4 Gas Chromatography . . . . . . . . . . . . . . . . . . . . . . 3859.9.5 Miscellaneous . . . . . . . . . . . . . . . . . . . . . . . . . . 3859.10 Cationic Surfactants . . . . . . . . . . . . . . . . . . . . . . . 386

Contents XIX

9.10.1 Titration Method . . . . . . . . . . . . . . . . . . . . . . . . 3869.10.2 Atomic Absorption Spectrometry (AAS) . . . . . . . . . . . . 3869.10.3 Gas Chromatography–Mass Spectrometry (GC–MS) . . . . . 3869.11 Anionic Surfactants . . . . . . . . . . . . . . . . . . . . . . . 3869.11.1 Titration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3869.11.2 Spectrophotometry . . . . . . . . . . . . . . . . . . . . . . . 3879.11.3 Atomic Absorption Spectrometry (AAS) . . . . . . . . . . . . 3879.11.4 High-Performance Liquid Chromatography (HPLC) . . . . . 3889.12 Non-Ionic Surfactants . . . . . . . . . . . . . . . . . . . . . . 3889.12.1 Spectrophotometry . . . . . . . . . . . . . . . . . . . . . . . 3889.12.2 Atomic Absorption Spectrometry (AAS) . . . . . . . . . . . . 3899.12.3 Liquid Chromatography–Mass Spectrometry (LC–MS) . . . . 3899.13 Aliphatic Chloro Compounds . . . . . . . . . . . . . . . . . . 3909.13.1 Gas Chromatography . . . . . . . . . . . . . . . . . . . . . . 3909.13.2 Purge and Trap Analysis . . . . . . . . . . . . . . . . . . . . 3909.13.3 Head Space Analysis . . . . . . . . . . . . . . . . . . . . . . . 3919.13.4 Miscellaneous . . . . . . . . . . . . . . . . . . . . . . . . . . 3929.14 Volatile Organic Compounds . . . . . . . . . . . . . . . . . . 3929.14.1 Head Space Analysis . . . . . . . . . . . . . . . . . . . . . . . 3929.14.2 Stripping Methods . . . . . . . . . . . . . . . . . . . . . . . . 3939.14.3 Mass Spectrometry . . . . . . . . . . . . . . . . . . . . . . . 3939.15 Chlorinated Dioxins . . . . . . . . . . . . . . . . . . . . . . . 3939.16 Nitrogen Compounds . . . . . . . . . . . . . . . . . . . . . . 3939.16.1 Spectrofluorometry . . . . . . . . . . . . . . . . . . . . . . . 3949.16.2 Proteins and Peptides . . . . . . . . . . . . . . . . . . . . . . 3979.16.3 Nucleic Acids . . . . . . . . . . . . . . . . . . . . . . . . . . 3979.16.4 Enzyme Activity . . . . . . . . . . . . . . . . . . . . . . . . . 3989.16.5 Aliphatic and Aromatic Amines . . . . . . . . . . . . . . . . 3989.16.6 Nitro-Compounds . . . . . . . . . . . . . . . . . . . . . . . . 3999.16.7 Azarenes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4009.16.8 Urea . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4009.16.9 Hydroxylamine . . . . . . . . . . . . . . . . . . . . . . . . . 4009.16.10 Acrylamide . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4009.16.11 Ethylene Diamine Tetracetic Acid and Nitriloacetic Acid . . . 4019.17 Sulfur Compounds . . . . . . . . . . . . . . . . . . . . . . . 4019.17.1 Alkyl Sulfides and Disulfides . . . . . . . . . . . . . . . . . . 4019.17.2 Thiols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4029.17.3 Dimethyl Sulfoxide . . . . . . . . . . . . . . . . . . . . . . . 4029.17.4 Thiabend Azole . . . . . . . . . . . . . . . . . . . . . . . . . 4029.17.5 Cysteine and Cystine . . . . . . . . . . . . . . . . . . . . . . 4039.17.6 Miscellaneous . . . . . . . . . . . . . . . . . . . . . . . . . . 4039.18 Chlorinated Insecticides . . . . . . . . . . . . . . . . . . . . 4039.18.1 Gas Chromatography . . . . . . . . . . . . . . . . . . . . . . 4039.18.2 High-Performance Liquid Chromatography . . . . . . . . . . 404

XX Contents

9.19 Polychlorobiphenyls . . . . . . . . . . . . . . . . . . . . . . . 4049.19.1 Gas Spectrofluorometry . . . . . . . . . . . . . . . . . . . . . 4059.19.2 Gas Chromatography . . . . . . . . . . . . . . . . . . . . . . 4059.19.3 Column Chromatography . . . . . . . . . . . . . . . . . . . . 4089.19.4 Miscellaneous . . . . . . . . . . . . . . . . . . . . . . . . . . 4099.20 Organophosphorus Compounds . . . . . . . . . . . . . . . . 4099.20.1 Spectrophotometric Method . . . . . . . . . . . . . . . . . . 4099.20.2 Gas Chromatography . . . . . . . . . . . . . . . . . . . . . . 4109.20.3 Enzymatic Methods . . . . . . . . . . . . . . . . . . . . . . . 4109.20.4 X-ray Fluorescence Spectrometry . . . . . . . . . . . . . . . . 4119.21 Azine Herbicides . . . . . . . . . . . . . . . . . . . . . . . . 4119.21.1 Gas Chromatography . . . . . . . . . . . . . . . . . . . . . . 4119.21.2 Gas Chromatography–Mass Spectrometry (GC–MS) . . . . . 4119.21.3 High-Performance Liquid Chromatography (HPLC) . . . . . 4119.22 Diuron, Irgalol, Chlorothalonil . . . . . . . . . . . . . . . . . 4129.23 Lipids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4129.24 Sterols . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4139.25 Chelators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4159.26 Humic Materials and Plant Pigments . . . . . . . . . . . . . . 4169.27 Vitamins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4239.28 Cobalamin . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4239.29 Pectenotoxins . . . . . . . . . . . . . . . . . . . . . . . . . . 4239.30 Flavins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4269.31 Microcystine . . . . . . . . . . . . . . . . . . . . . . . . . . . 4269.32 Preconcentration of Organics . . . . . . . . . . . . . . . . . . 426

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . 426

10 Organometallic Compounds . . . . . . . . . . . . . . . . . . . . 44310.1 Organoarsenic Compounds . . . . . . . . . . . . . . . . . . . 44310.1.1 Atomic Absorption Spectrometry . . . . . . . . . . . . . . . 44410.1.2 Spectrophotometric Method . . . . . . . . . . . . . . . . . . 44510.1.3 Miscellaneous Methods . . . . . . . . . . . . . . . . . . . . . 44610.2 Organocadmium Compounds . . . . . . . . . . . . . . . . . 44610.2.1 Anodic Scanning Voltammetry . . . . . . . . . . . . . . . . . 44610.3 Organocopper Compounds . . . . . . . . . . . . . . . . . . . 44610.4 Organolead Compounds . . . . . . . . . . . . . . . . . . . . 44710.5 Organomercury Compounds . . . . . . . . . . . . . . . . . . 44710.5.1 Atomic Absorption Spectrometry . . . . . . . . . . . . . . . 45010.5.2 Gas Chromatography . . . . . . . . . . . . . . . . . . . . . . 45210.5.3 Miscellaneous . . . . . . . . . . . . . . . . . . . . . . . . . . 45410.6 Organothallium Compounds . . . . . . . . . . . . . . . . . . 45410.7 Organotin Compounds . . . . . . . . . . . . . . . . . . . . . 45510.7.1 Atomic Absorption Spectrometry . . . . . . . . . . . . . . . 45510.7.2 Gas Chromatography . . . . . . . . . . . . . . . . . . . . . . 456

Contents XXI

10.7.3 HydrideGenerationGasChromatography–Microwave InducedAtomic Emission Spectrometry (HGGC–MIAES) . . . . . . . 459

10.7.4 Thermal Desorption-Gas Chromatography–InductivelyCoupled Plasma Mass Spectrometry (TDGC–ICPMS) . . . . . 460

10.7.5 High-Performance Liquid Chromatography . . . . . . . . . . 46110.7.6 Miscellaneous . . . . . . . . . . . . . . . . . . . . . . . . . . 461

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . 462

11 Elemental Analysis . . . . . . . . . . . . . . . . . . . . . . . . 46711.1 Boron . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46711.2 Total Iodine . . . . . . . . . . . . . . . . . . . . . . . . . . . 46711.3 Organic Nitrogen . . . . . . . . . . . . . . . . . . . . . . . . 46811.4 Organic Phosphorus . . . . . . . . . . . . . . . . . . . . . . . 47011.5 Silicon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47111.6 Total Sulfur . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47111.7 Carbon Functions . . . . . . . . . . . . . . . . . . . . . . . . 47211.7.1 Dissolved Organic Carbon . . . . . . . . . . . . . . . . . . . 47211.7.2 Dissolved Inorganic Carbon . . . . . . . . . . . . . . . . . . 48711.7.3 Particulate Organic Carbon . . . . . . . . . . . . . . . . . . . 48911.7.4 Dissolved Organic Carbon . . . . . . . . . . . . . . . . . . . 49011.7.5 Chemical Oxygen Demand . . . . . . . . . . . . . . . . . . . 49311.7.6 Biochemical Demand . . . . . . . . . . . . . . . . . . . . . . 49611.8 Oxygen Isotopic Ratios . . . . . . . . . . . . . . . . . . . . . 498

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . 498

Subject Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 505

1 Sampling and Storage

1.1Sampling

Common to all analytical methods is the need for correct sampling. It is stillthe most critical stage with respect of risks to accuracy in aquatic trace metalchemistry, owing to the potential introduction of contamination. Systematicerrors introduced here will make the analysis unreliable. Very severe errorswere commonly made during sampling by most laboratories until a decadeago, owing to ignorance or at least underestimation of the problems connectedwith sampling. This is the principal reason why nearly all trace metal databefore about 1975 for the sea and many fresh water systems are to be regardedas inaccurate or at least doubtful.

Surface-water samples are usually collected manually in precleaned poly-ethylene bottles (from a rubber or plastic boat) from the sea, lakes, and rivers.Sample collection is performed in the front of the bow of boats, against thewind. In the sea, or in larger inland lakes, sufficient distance (about 500 m) inan appropriate wind direction has to be kept between the boat and the researchvessel to avoid contamination. The collection of surface water samples from thevessel itself is impossible, considering the heavy metal contamination plumesurrounding each ship. Surface water samples are usually taken at 0.3–1 mdepth, in order to be representive and to avoid interference by the air/waterinterfacial layer in which organics and consequently bound heavy metals accu-mulate. Usually, sample volumes between 0.5 and 2 l are collected. Substantiallylarger volumes could not be handled in a sufficiently contamination-free man-ner in subsequent sample pretreatment steps.

Reliable deep-water sampling is a special and demanding art. It usually hasto be done from the research vessel. Special devices and techniques have beendeveloped to provide reliable samples.

Samples for mercury analysis should preferably be taken in pre-cleanedflasks. If, as required for the other ecotoxic heavy metals, polyethylene flasksare commonly used for sampling, then an aliquot of the collected water samplefor the mercury determination has to be transferred as soon as possible intoglass bottles, because mercury losses with time are to be expected in polyeth-ylene bottles.

2 1 Sampling and Storage

Ashton and Chan [1] have reviewed the techniques for the collection ofseawater samples: preservation, storage, and prevention of contamination areall discussed. The most appropriate measurement techniques, preconcentra-tion and extraction, method validation, and analytical control are all covered.The apparent aluminium content of seawater stored in ordinary containerssuch as glass and polyethylene bottles decreases gradually, e.g., to half in 2.5 h.But if the samples are acidified with 0.5 ml/l concentrated sulfuric acid thealuminium content remains constant for at least one month. Accordingly, sam-ples should be acidified immediately after collection. However, the aluminiumcould be recovered by acidifying the stored samples and leaving them for atleast five hours.

Shipboard analysis for the sampling of trace metals in seawater has beendiscussed by Schuessler and Kremling [2] and Dunn et al. [3]. Teasdale et al.have reviewed methods for collection of sediment pore-waters using in situdialysis samples [4]. Bufflap and Allen [5] compared centrifugation, squeezing,vacuum filtration, and dialysis methods for sediment pore-water sampling.

Yamamoto et al. [6] studied preservation of arsenic- and antimony-bearingsamples of seawater. One-half of the sample (20 l) was acidified to pH 1 withhydrochloric acid immediately after sampling, and the remaining half was keptwithout acidification. In order to clarify the effect of acidification on storage,measurements were made over a period of a month after sampling. Results aregiven in Table 1.1. In this study, a standard addition method and calibrationcurve method were used for comparison and it was proven that the two gavethe same results for the analyses of seawater.

Table 1.1. Effects of storage on the concentration of arsenic and antimony in seawater(mg/l) [6]

Day As (total) As (III) Sb (total) Sb (III)

1 0.75 0.30 0.21 0(0.66)∗ (0.41) (0.21) (0)

2 0.63 0.25 0.21 0(0.73) (0.65) (0.23) (0)

7 0.70 0.35 0.25 0(0.75) (0.42) (0.23) (0)

14 0.75 0.19 0.22 0(0.60) (0.39) (0.23) (0)

30 0.73 0.35 0.19 0(0.68) (0.41) (0.20) (0)

Average 0.72 ± 0.05 0.27 ± 0.07 0.22 ± 0.22 0(0.70 ± 0.07) (0.41 ± 0.01) (0.22 ± 0.01) (0)

∗Values in parentheses were those for the non-acidified sample.

1.2 Sampling Devices 3

Yamamoto et al. [6] conclude that their method was quite successful forthe species-specific determination of arsenic and antimony in seawater. Thesemethods, especially those for the determination of arsenic (III) and anti-mony (III), are quite satisfactory, as the method is almost free from interferenceof foreign ions.

The total concentrations of arsenic and antimony may be underestimated,because organic species of these elements may have been overlooked. Thereremains ambiguity in defining the difference in concentration of arsenic (total)and arsenic (III) to arsenic (V), and that of antimony (total) and antimony (III)to antimony (V).

1.2Sampling Devices

The job of the analyst begins with the taking of the sample. The choice ofsampling gear can often determine the validity of the sample taken; if con-tamination is introduced in the sampling process itself, no amount of care inthe analysis can save the results. Sampling devices and sample handling forhydrocarbon analysis have been exhaustively reviewed by Green [7].

The highest concentrations of naturally occurring dissolved and particulateorganic matter in the oceans are normally found in the surface films. Whenorganic pollutants are present, they too, tend to accumulate in this surface film,particularly if they are either non-polar or surface active. Much of the availableinformation on these surface films is reviewed by Wangersky [8, 9].

A major problem in the sampling of surface films is the inclusion of water inthe film. In the ideal sampler, only the film of organic molecules, perhaps a fewmolecular layers in thickness, floating on the water surface, would be removed;the analytical results should then be expressed either in terms of volume takenor of surface area sampled.

In practice, all of the samplers in common use at this time collect someportion of the surface layer of water. Each sampler collects a different slice ofthe surface layer; thus results expressed in weights per unit volume are onlycompared when the samples were taken with the same sampler. Of the surfacefilm samplers described in the literature, the “slurp” bottle used by Gordonand Keizer [10] takes perhaps the deepest cut, sampling as far as 3 mm into thewater. The sampling apparatus consists simply of an evacuated container fromwhich extend floating tubes with holes. The main advantage of the samplerare its simplicity and low cost. Its disadvantages are the thickness of the waterlayer sampled and the inherent inability to translate the results of the analysisinto units of weight per surface area swept.

A thinner and more uniform slice of the surface can be collected with theHarvey [11] rotating drum sampler. This consists of a floating drum whichis rotated by an electric motor, the film adhering to the drum being removedby a windshield wiper blade. The thickness of the layer sampled depends


upon the coating on the drum, the speed with which it rotates, and the watertemperature. The slice taken will usually run between 60 and 100 µm. Thismethod has the advantage of sampling a known area of ocean surface, and ofcollecting a large sample very quickly. However, like most of the other surfacefilm samplers, it must be operated from a small boat and can therefore only beused in calm weather.

The collector most commonly used for the surface films is the Garrettscreen [12], a monel metal or stainless steel screen which is dipped verticallybelow the surface and then raised in a horizontal position through the surfacefilm. The material clinging to the screen is then drained into a collection bottle.This sampler collects a slice of surface somewhere between 150 and 450 µmthick. A relatively small sample, about 20 ml, is collected on each dip, so that thecollection of a sample of reasonable size is very time-consuming. This methodis also limited to calm weather, since the sampling must be conducted froma small boat. While the Garrett screen has been adopted by many investigatorsbecause of its simplicity and low cost, as well as its relative freedom fromcontamination, it is a far from satisfactory solution to the problem of surface-film sampling. The small size of the sample taken greatly restricts the kinds ofinformationwhich canbe extracted fromthe sample.Also, theuncertainty as tothe thicknessof thefilmsamplesmakes comparisonbetween samplersdifficult.

Only one actual comparison of these techniques is available in the literature.Daumas et al. [13] compared the Harvey drum sampler to the Garrett screen,and found greater organic enrichment in both dissolved and particulate matterin the drum samples. The size of the difference between the two samplerssuggests that the Garrett screen included 2–3 times as much water as did theHarvey drum.

Several samplers have been constructed on the principle of the use of a spe-cially treatedsurface tocollect surface-activematerials.HarveyandBurzell [14]used a glass plate, which was inserted and removed vertically with the materialadhering to the plate then collected with a wiper. This method of sampling stillincludes some water; if a surface which preferentially absorbs hydrophobic orsurface-active material is used, for example, Teflon, either normal or speciallytreated, only the organic materials will be removed, and the water will drainaway. Such samplers have been described by Garrett and Burger [15], Larssonet al. [16] and Miget et al. [17], amongst others. Anufrieva et al. [18] usedpolyurethane sheets, rather than Teflon, as the absorbent. While this sort ofsampler seems, at least theoretically, to have many advantages, since the surfaceswept is easily defined and the underlying water is excluded, the sample takenis very small. In addition, it must be removed from the sampler by elution withan organic solvent, so the chances of contamination from the reagents usedare sharply increased.

An interesting method which combines sampling and analysis in one stephas been described by Baier [19]. A germanium probe is dipped into the waterand carefully withdrawn, bringing with it a layer of surface-active material.


This layer is then analysed directly by internal reflectance infra-red spec-troscopy. Since there is no handling of the sample, contamination is reducedto a minimum. However, only infra-red spectral analysis is possible with thissystem; since the material absorbed on the germanium prism is always a mix-ture of compounds, and since the spectrophotometer used for the productionof the spectra is not a high-precision unit, the information coming from thistechnique is limited. While identification of specific compounds is not usuallypossible, changes in spectra, which can be related to the time of day, season,or to singular events, can be observed.

Overall there is still no really satisfactory method for sampling the surfacefilm. Of the methods now in use, the method favoured by most workers isfavoured for practical reasons and not for any inherent superiority as a collec-tor. Also, as long as we have no simple accurate method for measuring totaldissolved organic carbon, it will be difficult to estimate the efficiency of anysurface film collectors.

Sampling the subsurface waters, although simpler than sampling the sur-face film, also presents some not completely obvious problems. For example,the material from which the sampler is constructed must not add any or-ganic matter to the sample. To be completely safe, then the sampler shouldbe constructed either of glass or of metal. All-glass samplers have been usedsuccessfully at shallow depths; these samplers are generally not commerciallyavailable [20, 21]. To avoid contamination from material in the surface film,these samplers are often designed to be closed while they are lowered throughthe surface, and then opened at the depth of sampling. The pressure differentiallimits the depth of sampling to the upper 100 m; below this depth, implosionof the sampler becomes a problem.

Implosion at greater depth can be prevented either by strengthening thecontainer or by supplying pressure compensation. The first solution has beenapplied in the Blumer sampler [22]. The glass container is actually a linerinside an aluminium pressure housing; the evacuated sampler is lowered tothe required depth, where a rupture disc breaks, allowing the sampler to fill.Even with the aluminium pressure casing, however, the sampler cannot be usedbelow a few thousand metres without damage to the glass liner.

Another approach to the construction of glass sampling containers involvesequalisation of pressure during the lowering of the sampler. Such a samplerhas been described by Bertoni and Melchiorri-Santolini [23]. Gas pressure issupplied by a standard diver’s gas cylinder, through an automatic delivery valveof the type used by SCUBA divers. When the sampler is opened to the water,the pressurising gas is allowed to flow out as the water flows in. The samplerin its original form was designed for use in Lago Maggiore, Italy, where themaximum depth is about 200 m, but in principal it can be built to operate atany depth.

Stainless steel samplers have been devised, largely to prevent organic con-tamination. Some have been produced commercially. The Bodega–Bodman


sampler and the stainless steel Niskin bottle, formerly manufactured by Gen-eral Oceanics, Inc., are examples. These bottles are both heavy and expensive.The Bodega–Bodman bottle, designed to take very large samples, can only beattached to the bottom of the sampling wire; therefore, the number of samplestaken on a single station is limited by the wire time available, and depth profilesrequire a great deal of station time.

The limitations of glass and stainless steel samplers have led many workersto use the more readily available plastic samplers, sometimes with the fullknowledge of the risks and sometimes with the hope that the effects resultingfrom the choice of sampler will be small compared with the amounts of organicmatter present. The effects of the containers can be of three classes:

1. Organic materials may be contributed to the sample, usually from the plas-ticisers used in the manufacture of the samplers.

2. Organic materials, particularly hydrophobic compounds, may be absorbedfrom solution on the walls of the sampler.

3. Organic materials may be absorbed from the surface film or surface waters,then desorbed into the water samples at depth, thereby smearing the realvertical distributions.

The first case is the most likely to be a problem with new plastic samplers.Although there is little in the literature to substantiate the belief, folklore has itthat aging most plastic samplers in seawater markedly reduces the subsequentleaching of plasticisers. The second case is known to be a problem; in fact,the effect is used in the various Teflon surface film samplers already men-tioned. This problem alone would seem to militate against the use of Teflonfor any sampling of organic materials, unless a solvent wash of the sampler inincluded routinely. With such a solvent wash, we introduce all of the problemsof impurities in the reagents.

The third, and largelyunexpected, case appearedas aproblem in theanalysisof petroleum hydrocarbons in seawater [24]. In this case, petroleum hydrocar-bons, picked up presumably in the surface layers or surface film, were carrieddown by the sampling bottles and were measured as part of the pollutant load ofthe deeper waters. While the possibility of absorption and subsequent releaseis obviously most acute with hydrophobic compounds and plastic samplers, itdoes raise a question as to whether any form of sampler which is open on itspassage through the water column can be used for the collection of surface-active materials. The effects of such transfer of material may be unimportant inthe analysis of total organic carbon, but could be a major factor in the analysisof single compounds or classes of compounds.

Again, as in the case of the surface film samplers, information on the com-parative merits of the various water samplers is largely anecdotal. Althoughsuch studies are not inspiring and require an inordinate amount of time, bothon the hydrographical wire and in the laboratory, they are as necessary fora proper interpretation of data as are intercalibration studies of the analytical


methods. The lack of comparison studies of the various samplers increases theprobability of polemics in the literature.

Smith [25] has described a device for sampling immediately above thesediment water interface of the ocean. The device consists of a nozzle supportedby a benthic sled, a hose, and a centrifugal deck pump, and is operated froma floating platform. Water immediately above the sediment surface is drawnthrough the nozzle and pumped through the hose to the floating platform,where samples are taken. The benthic sled is manipulated by means of a handwinch and a hydrowire.

1.2.1Intercomparison of Seawater Sampling for Trace Metals

Several round-robin intercalibrations for trace metals in seawater [26–30] havedemonstrated a marked improvement in both analytical precision and numer-ical agreement of results among different laboratories. However, it has oftenbeen claimed that spurious results for the determination of metals in seawatercan arise unless certain sampling devices and practical methods of samplerdeployment are applied to the collection of seawater samples. It is thereforedesirable that the biases arising through the use of different, commonly usedsampling techniques be assessed to decide upon the most appropriate tech-nique(s) for both oceanic baseline and nearshore pollution studies.

Two international organisations, the International Council for the Explo-ration of the Sea (ICES) and the Intergovernmental Oceanographic Commis-sion (IOC), have sponsored activities aimed at improving the determinationof trace constituents in seawater through intercalibrations. Since 1975, ICEShas conducted a series of trace metal intercalibrations to assess the compa-rability of data from a number of laboratories. These exercises have includedthe analysis of both standard solutions and real seawater samples [26–31]. Theconsiderable improvement in the precision and relevant agreement betweenlaboratories has been reflected in the results of these intercalibrations. By 1979it had been concluded that sufficient laboratories were capable of conductinghigh-precision analyses of seawater for several metals to allow an examinationto be made of the difference between commonly used sampling techniques forseawater sample collection.

In early 1980, the IOC, with the support of the World Meteorological Or-ganisation (WMO) and the United Nations Environment Program (UNEP),organised a workshop on the intercalibration of sampling procedures at theBermuda Biological Station, during which the most commonly used samplingbottles and hydrowires were to be inter-compared. This exercise forms partof the IOC/WMO/UNEP Pilot Project on monitoring background levels of se-lected pollutants in open-ocean waters. Windom [32] had already conducteda surveyof the seawater samplingandanalytical techniquesusedbymarine lab-oratories, and the conclusions of the survey were largely used for the selection


of sampling devices to be intercompared. The bottles selected for compari-son in Bermuda were modified and unmodified GO-FLO® samplers, modifiedNiskin® bottles, and unmodified Hydro-Bios® bottles. GO-FLO samplers arethe most widely used sampling device for trace metals in seawater. The othertwo devices continue to be used by several marine laboratories. Windom’s [32]1979 survey established that the most common method of sampler deploymentwas on hydrowires, as opposed to the use of rosette systems. The hydrowiresselected for intercomparison were Kevlar®, stainless steel, and plastic-coatedsteel. Kevlar and plastic-coated steel were selected because they are widelyused in continental shelf and nearshore environments, and are believed to berelevantly “clean”.

The method of intercomparison of the various devices was to deploy pairsof sampler types on different hydrowires to collect water samples from a ho-mogeneous body of deep water at Ocean Station S (Panulirus Station) nearBermuda (Fig. 1.1). The water at this depth has characteristics of 3.97±0.05 ◦C

Figure 1.1. Sampling strat-egy. From [29]


temperature and 35.01 ± 0.02% salinity for the month of January [33]. Therestricted length of Kevlar hydrowire available necessitated the collection ofsamples in the lower thermocline at depths between 1150 and 1250 m.

Data analysis was reduced to a separate one-way analysis of variance on thedata from individual laboratories in order to examine the difference betweentypes of sampling bottle on a single (common) hydrowire, and to determinethe influences of the three types of hydrowire using a single type of samplingbottle (modified GO-FLO). Samples were replicated so that there were, in allcases, two or more replicates to determine the lowest level and analytical error.

Replicate unfiltrated water samples were collected for each participant forthe comparison of pairs of sampling bottles on different hydrowires [29].Modified GO-FLO bottles were employed on each of the three hydrowires,

Table 1.2. Numerical comparisons for nickel (µg/l)∗ [29]

WireBottle

PCSHB

PCSMGF

SSMod GF

SSExw GF

KEVNIS

KEVMGF

PCSMGF

SSMGF

KEVMGF

Laboratory1 m sd 0.224 0.205 0.209 0.243 0.233 0.207 0.201 0.226 0.207

0.015 0.020 0.023 0.013 0.023 0.025 0.029 0.025 0.0252 m sd 0.298 0.278 0.235 0.235 0.218 0.150 0.223 0.235 0.231

0.031 0.123 0.077 0.048 0.019 0.018 0.128 0.059 0.1194 m sd 0.18 0.47 0.47 0.35 0.47 0.41 0.23

0.07 – – 0.17 – 0.12 –5 m sd 0.478 0.221 0.240 0.235 0.273 0.237 0.193 0.238 0.237

0.066 0.026 0.014 0.012 0.021 0.016 0.048 0.012 0.0166 m sd 0.340 0.159 0.220 0.238 0.237 0.232 0.159 0.230 0.230

0.052 0.035 0.030 0.021 0.023 0.026 0.035 0.027 0.0247 m sd 1.93 163 1.60 1.75 1.63 1.68 1.72

0.24 0.10 0.37 0.13 0.10 0.27 0.258A m sd 0.185 0.100 0.123 0.113 0.160 0.105 0.100 0.119 0.123

0.041 – 0.005 0.012 0.54 0.010 – 0.009 0.04110 m sd 0.737 0.357 0.3634 0.353 0.385 0.461 0.413 0.493 0.462

0.078 0.77 0309 0.131 0.064 0.162 0.159 0.262 0.16211 m sd 0.511 0.367 0.349 0.365 0.421 0.393 0.367 0.357 0.404

0.034 0.008 0.034 0.009 0.009 0.027 0.008 0.025 0.03712 m sd 0.230 0.165

0.024 0.03613 m sd 0.230 0.200 0.238 0.265 0.204 0.236 0.200 0.250 0.236

0.036 0.020 0.045 0.007 0.056 0.007 0.020 0.035 0.007

∗Numbers result fromcommoncomputeranalyses andnotall suchfigureswill benecessarilysignificant; PCS = Plastic-coated steel hydrowire; SS = Stainless steel (type 302 unlubri-cated) hydrowire; KEV = Kevlar® hydrowire; HB = Hydro-bios sampler; MGF = ModifiedGO-FLO sampler; Mod GF = Unmodified GO-FLO sampler; Exw GF = Unmodified GO-FLO sampler; NIS = Modified Niskin sampler; m = mean; sd = standard deviation


and this permitted a comparison of the three types of hydrowire. Only inthe cases of iron and manganese were there indications of inhomogeneity atlevels that might invalidate the intercomparison. This is assumed to be due toinhomogeneity in the distribution of suspended particulate material, whichwill influence metals that have major fractions in the particulate phase.

The results obtained by various calibrations in the determination of nickeland copper are shown in Tables 1.2 and 1.3. Table 1.4 gives the differencesbetween sampling devices for copper, as determined by each participant, whenthese are significant at the 95% and 90% levels of confidence. Only the resultsof participants that had acceptable analytical performance, as measured byprecision and agreement with contemporary consensus values for deep NorthAtlantic waters (Table 1.5), were used for drawing conclusions.

The experiment reveals that the differences between results obtainedthrough the use of various combinations of hydrowires and samplers are notlarge, and in no case can they account for the recent decline in the ocean

Table 1.3. Statistical comparisons for nickel [29]

Basecomparison

PCSMGF/MGF

PCSHB/MGF

SSMGF/MG

KEVMGF/MGF

KEVNIS/MGF

MGFWIRES

Laboratory1 NS Sig

HB>MGFSigGF>MGF

NS SigNIS>MGF

SigSS>KEV>PCS

2 NS NS NS Sig SigNIS>MGF

NS

4 Sig NS NS5 Sig

HB>MGFNS 90

NIS>MGFSigSS>KEV>PCS

6 SigHB>MGF

SigGF>MGF

NS NS SigSS>KEV>PCS

8A SigHB>MGF

10 NS SigHB>MGF

NS Sig NS NS

11 SigHB>MGF

NS NS NS SigSS>KEV>PCS

12 90 SigHB>MGF

13 NS NS NS NS 90SS>KEV>PCS

PCS = Plastic-coated steel hydrowire; SS = Stainless steel (type 302 unlubricated)hydrowire; KEV = Kevlar® hydrowire; HB = Hydro-bios sampler; MGF = ModifiedGO-FLO sampler; GF = Unmodified GO-FLO sampler; NIS = Modified Niskin sampler;Sig = Difference is significant (P < 0.05); 90 = Difference is significant (P < 0.01);NS = Not significant (P < 0.01)


Table 1.4. Numerical comparisons for copper (µg/l) [29]

Wire:Bottle:

PCSHB

PCSMGF

SSMod GF

SSExw GF

KEVNIS

KEVMGF

PCSMGF

SSMGF

KEVMGF

1 msd

0.0940.007

0.0920.009

0.0950.012

0.1030.012

0.1110.011

0.1310.012

0.0930.011

0.0990.012

0.1200.021

2 msd

1.0000.857

0.7650.289

0.5530.261

0.6200.100

0.4550.487

0.2720.185

0.6500.291

0.5860.186

0.4030.298

3 msd

0.4370.347

0.1800.084

0.110.067

0.2050.034

0.4470.540

0.5500.317

0.2330.081

0.2080.051

0.4550.314

4 msd

0.5330.163

0.4350.177

1.250.35

1.0650.177

0.4350.177

1.1580.252

1.0650.177

5 msd

0.1880.108

0.0630.003

0.0640.004

0.1420.010

0.1010.049

0.0720.012

0.1010.059

0.1030.043

0.0720.012

6 msd

0.615†

0.5600.0700.005

0.0740.003

0.0830.004

0.1210.039

0.1200.022

0.0700.005

0.0790.006

0.1200.026

7 msd

0.350.29

0.270.12

0.710.59

0.280.23

0.270.12

0.500.47

0.320.37

8B msd

0.1550.076

0.0450.006

0.1630.044

0.2780.059

0.1330.030

0.1600.037

0.0450.006

0.2200.078

0.1400.038

9 msd

0.840.79

0.320.03

0.350.02

0.550.21

0.320.03

0.440.17

10 msd

0.1230.015

0.1350.003

0.1580.033

0.1190.032

0.0960.015

0.1000.019

0.1300.024

0.1380.037

0.1010.019

11 msd

0.1950.089

0.1370.027

0.1020.005

0.1060.001

0.1090.013

0.1320.019

0.1370.027

0.1040.004

0.1490.073

12 msd

0.059†

0.3250.1720.040

13 msd

0.1680.063

0.1010.028

0.1050.013

0.2920.006

0.1330.009

0.1210.020

0.1010.028

0.1860.100

0.1210.200

Numbers result from common computer analyses and not all such figures will be necessarilysignificant. † Suspected contamination. All other symbols are the same as those used inTable 1.2.

concentrations of trace metals reported in the literature. Nevertheless, for sev-eral metals, most notably copper, nickel, and zinc, significant differences areevident between both bottles and hydrowires. For deep ocean studies the bestcombination of those tested is undoubtedly modified GO-FLO samplers andplastic-coated steel hydrowire. Except in the cases of mercury and manganese,Hydro-Bios samplers appear to yield higher values than modified GO-FLOsamplers. In contrast, Niskin bottles, modified by the replacement of the inter-nal spring by silicone tubing, are capable of collecting samples of comparablequality to those collected by modified GO-FLO samplers for all metals exceptzinc. Modification to factory-supplied Teflon GO-FLO bottles, (i.e., replace-ment of O-rings with silicone equivalents and the substitution of all Teflondrain cocks for those originally supplied), do appear to result in a significant


Table 1.5. Results of sampling bottle and hydrowire intercomparison [29]

Metal Concentration No of Best combined ComparisonsLaboratories sampling/ Hydrowires Samplers

analyticalprecisions(µg/l)

Cd 0.035 ± 0.016 12 0.001 PCS<(KEV ≈ SS) (MGF ≈NIS)<HB<GF

Cu 0.13 ± 0.040.51 ± 0.28

66

0.010 PCS<(KEV ≈ SS) (MGF ≈NIS)<HB<GF

Ni 0.21 ± 0.050.42 ± 0.11

73

0.02 PCS<(KEV ≈ SS) (MGF ≈NIS ≈ GF)<HB

Zn 0.35 ± 0.18 5 0.05 PCS<(KEV ≈ SS) MGF<(NIS ≈HB<≈ GF)

Fe 0.41 ± 0.29 3 0.05 PCS<(KEV<SS) (MGF ≈NIS)<GF<HB

Mn 0.064 ± 0.0380.012 ± 0.006

21

0.0100.003

(PCS ≈ KEV ≈SS)

(MGF ≈ NIS ≈GF ≈ HB)

Hg 0.007 ± 0.0020.001

21

0.002 Insufficient comparisons

∗Mean (±SD, µg/l); Other symbols the same as used in Table 1.4.

reduction in the levels of most metals in seawater samples collected with them.Kevlar and stainless steel hydrowires generally yield measurably greater con-centrations of those metals than does plastic-coated steel. These differences,however, are small enough to suggest that these hydrowires are still suitablefor trace metal studies of all but the most metal-depleted waters if properprecautions are taken [34–37].

A major conclusion of the Bermuda experiment is that the use of differingsampling devices and hydrowires only accounts for a small portion of thedifferences between trace metal results from different laboratories. It appearsthat the major contributions to such differences are analytical artefacts. It isstressed that although the sampling tools available to the marine geochemistappear adequate for the measurement of metal distribution in the ocean, theexecution of cooperative monitoring programs for metals should be precededby mandatory intercomparison of sample storage and analytical procedures.

1.2.2Intercomparison of Sampling Devices and Analytical Techniques Using Seawaterfrom a CEPEX (Controlled, Ecosystem Pollution Experiment) Enclosure

Wong et al. [38] conducted an intercomparison of sampling devices usingseawater at 9 m in a plastic enclosure of 65 m in Saanich Inlet, BC, Canada. Thesampling methods were:


1. Peristaltic pumping with Teflon tubing2. Niskin polyvinyl chloride (PVC) sampler3. Go-Flo sampler4. close–open–close sampler5. Teflon-piston sampler

Sampling was conducted for four days:

1. Day 1 (August 1978) for mercury.2. Day 2 for lead, cadmium, copper, cobalt, and nickel by Chelex extraction

and differential pulse polarography, as well as manganese by Chelex andflameless atomic absorptiometry.

3. Day 3 for lead by isotope dilution.4. Day 4 for cadmium, copper, iron, lead, and zinc by Freon extraction and

flameless atomic absorptiometry.

Samples were processed in clean rooms in the shore laboratory within 30 minof sampling. Results indicated (i) the feasibility of inter-calibrating using theenclosure approach; (ii) the availability of chemical techniques of sufficientprecision in the case of copper, nickel, lead, and cobalt for sampler inter-comparison and storage tests; and (iii) a problem in sub-sampling from thecaptured seawater in a sampler, and the difficulty of commonly used samplersto sample seawater in an uncontaminated way at the desired depth.

The Teflon tubing used in the pumping system, the Niskin sampler andthe Go-Flo sampler were cleaned by immersion in 0.05% nitric acid for thetubing and by soaking the inside of the samplers in 0.05% nitric acid overnight,rinsing with distilled water and repeating the dilute acid/distilled water cycle.The close–open–close sampler was cleaned by 0.1 N nitric acid overnight, thenrinsed with distilled water until the blank was acceptable. The Teflon-pistonsampler was cleaned by sucking in 0.05% nitric acid and standing overnight(in the case of the polythene bag liner used in the Teflon-piston sampler,hydrochloric acid was used instead of nitric acid).

The storage bottles were cleaned as follows. The Pyrex bottles (2 litres) wereused for mercury samples only. They were cleaned by filling with a solution of0.1% KMnO4, 0.1% K2S2O8, and 2% nitric acid, heated to 80 ◦C for 2 h, and aftercooling and rinsing, stored filled with 2% nitric acid containing 0.01% K2C2O7until ready for use. Conventional 1- or 2 l polyethylene bottles were used forthe other metal samples. They were cleaned by Patterson’s method [39]. Allbottles were stored inside two or three plastic bags to prevent contamination.

For the pumping system, seawater was pumped up from 9 m and collected inthe appropriate bottles on the raft and returned to the shore clean laboratoryfor preservation and/or analysis. For the other four sampling devices, thesampler was lowered to 9 m, allowed to equilibrate for 10 min, closed bytriggering mechanism activated by the Teflon messenger, raised to the surface,transferred into the container, transported back by boat and trucked back


to the shore clean laboratory, where samples were drawn. The time betweenmessenger activation and subsampling was about 30 min. For handling of thesamples, messengers, Teflon tubing, vinyl-coated hydrowires, and samplingdevices, all personnel wore polyethylene gloves to avoid contamination.

The clean laboratory for trace metals was divided into three areas: entrancelaboratory (with clothes changing annex), instrument laboratory, and ultra-clean sample preparation laboratory, all under positive pressure with activecharcoal filtered air. Personnel using the clean rooms were required to wearhair caps, polyethylene gloves, laboratory coats, and designated shoes. Theseitems are worn only in the clean rooms.

Mercury was determined after suitable digestion by the cold vapour atomicabsorption method [40]. Lead was determined after digestion by a stableisotope dilution technique [41–43]. Copper, lead, cadmium, nickel, and cobaltwere determined by differential pulse polarography following concentrationby Chelex 100 ion-exchange resin [44,45], and also by the Freon TF extractiontechnique [46]. Manganese was determined by flameless atomic absorptionspectrometry (FAA).

The precision of the procedures under clean room conditions is shown inTable 1.6.

The results in Fig. 1.2 show values between 0.06 and 0.12 nmol/kg. Theaverage mercury contents obtained by pumping, Niskin sampler, Go-Flo sam-pler, and the close–open–close device are 0.09 ± 0.01, 0.08 ± 0.03, and 0.10 ±0.02 nmol/kg, respectively.

The mercury values obtained by the Teflon-piston sampler were high at0.21 ± 0.2 nmol/kg due to malfunction with incomplete filling and previouscontamination, as indicated by the very low salinity in this set. The valuesinside the bag were higher than those outside, measured about one month afterintercomparison to be 0.02, 0.03, and 0.04 nmol/kg. There was a subsamplingproblem. The first and second draw of the sampling bottle usually showeda very wide spread of values, as much as 0.07 nmol/kg, e.g., between 0.05 and

Table 1.6. Precision of the procedure for Cu, Ni, Cd, Zn, Fe and Pb as applied in oceanchemistry clean room [38]

Metal Seawater Blank (nmol) Relative standard deviation Recovery (%)concentration at test level (average of 10(nmol/kg) analyses; %)

Cd 1.16 0.009 6 83–98Cu 13.0 0.08 2 95Ni 14.8 0.12 2 95Zn 32.3 0.69 2 95Fe 7.8 0.54 3 90Pb 0.10 0.02 30 85–100


Figure 1.2. Comparison of mercury concentrations found in a CEPEX enclosure using fivedifferent sampling methods. • = Pump; � = Hydrobios; � = Go-Flow; � = Niskin;� = Seakern. From [38]

0.12 nmol/kg.Thisdifferencewas real since the techniqueof coldvapouratomicabsorption should be capable of detecting differences in subsamples fromthe same digested sample in a Pyrex bottle. The peristaltic pumping methodappears to yield the best agreement between subsamples, with a differenceof 0.02, 0.00, and 0.01 nmol/kg between subsamples from the three casts. Theaverage mercury values for each sampler appeared to converge towards lowervalues on repeated casts within the same day. Further work is required to clarifycontamination in the mercury sampling.

Isotope dilution and mass spectrometry showed the lead values to be0.73 ± 0.02, 0.72 ± 0.03, 0.75 ± 0.02, 0.78 ± 0.05, and 0.81 ± 0.03 nmol/kg forsampling by peristaltic pump, Niskin sampler, Go-Flo sampler, close–open–close sampler, and Teflon-piston sampler, respectively (Fig. 1.3). For the othertwo techniques, the Teflon-piston sampler results showed considerable vari-ability and statistically much higher values. The results were not used in thecomparison. The Freon extraction and cold vapour atomic absorption ap-proach showed the same range of values as the isotope dilution approach, i.e.,0.071± 0.36, 0.76± 0.13, 0.73± 0.13 nmol/kg for the pumping, Niskin samplerand close–open–close sampler, respectively, with the exception of the Go-Flosampler with low value of 0.58 ± 0.15 nmol/kg. However, the range of valueswas wide, e.g., for the peristaltic pumping, 1.12 nmol/kg for the first cast, drop-ping to 0.46 nmol/kg for the third cast. Chelex extraction and differential pulsepolarography showed an even larger spread from 1.09 ± 0.26 nmol/kg for theclose–open–close sampler.


Figure 1.3. Comparison of lead concentrations found in a CEPEX enclosure using fivedifferent sampling methods and IDMS analysis. • = Pump;� = Hydrobios;� = Go-Flow;� = Niskin; � = Seakern. From [38]

Wong et al. [38] concluded that:

1. It is feasible to capture a large volume of seawater in the range of 65 000 lby the CEPEX approach for the purpose of sampler intercomparison. It ispossible by artificial stimulation of a plankton bloom and detritus removalto produce a reasonably homogeneous body of seawater for the study. Prox-imity of the in situ enclosure for the experiment and the on-shore, cleanlaboratory facilities eliminate errors introduced by shipboard contamina-tion under less than ideal conditions on cruises.

2. The following analytical techniques seem to be adequate for the concentra-tions under consideration: copper and nickel by Freon extraction and FAAcold vapour atomic absorption spectrometry, cobalt by Chelex extractionand differential pulse polarography, mercury by cold vapour atomic absorp-tion absorptiometry, lead by isotope dilution plus clean room manipulationand mass spectrometry. These techniques may be used to detect changes inthe above elements for storage tests: Cu at 8 nmol/kg, Ni at 5 nmol/kg, Co at0.5 nmol/kg, Hg at 0.1 nmol/kg, and Pb at 0.7 nmol/kg.

3. Salinity of seawater captured by various sampling devices in the CEPEXenclosure indicates problems not revealed in the usual oceanographic sam-pling situation. Relative to peristaltic pumping, all samples exhibited somesalinity anomalies. Inadequate flushing to rinse the sampler of any concen-trated brine or entrapped seawater is thought to be a problem.

4. Logistics and cleaning procedures are important factors in successful sam-pler intercomparison. It isnotdesirableorpossible toendorseor tocondemn

1.3 Sample Preservation and Storage 17

the performance of a certain type of sampler or analytical technique basedon results of one set of tests, especially if procedures are changed.

5. The problems of subsampling from the same seawater sample has to bestudied in greater detail.

6. A long-term but substained effort on sampler intercomparison would beadvantageous in identifying problems.

1.3Sample Preservation and Storage

If we were to choose the ideal method for the analysis of any component ofseawater, it would naturally be an in situ method. Where such a method ispossible, the problems of sampling and sample handling are eliminated and inmany cases we can obtain continuous profiles rather than limited number ofdiscrete samples. In the absence of an in situ method, the next most acceptablealternative is analysis on board ship. A “real-time” analysis not only permitsus to choose our next sampling station on the basis of the results of the laststation, it also avoids the problem of the storage of samples until the return toa shore laboratory.

While there are a few methods of this type for major constituents, and theadvent of the automatic analyser has made possible the adaptation of somemicro-methods to shipboard analysis, the majority of chemical analyses, par-ticularly those using the newer, more sophisticated instruments, must still berun on shore. For such samples, the problems of storage and sample preser-vation becomes all-important, since the quantity we wish to measure is the insitu value and not the amount remaining after some period of biological andchemical activity. Complete oxidation of the organic material to inorganic con-stituents takes longer than a year [47], however, decomposition great enoughto free most inorganic micronutrients takes place within the first two weeks.Important changes in micronutrient levels resulting from bacterial utilisationof organic compounds can be seen after one day. Therefore, some method ofpreservation of the organic compounds must be sought if the samples are tobe taken back to a shore laboratory.

Changes in the distribution of organic compounds in a seawater samplecan be due to physical, chemical, or biological factors. As a physical factor, wemight consider the absorption of surface-active materials on the walls of thesample container. While this effect cannot be eliminated it can be minimisedby the use of the largest convenient sample bottle, and the avoidance of plastic(especially Teflon) containers. Another possible method of eliminating thissource of error would be to draw the sample directly into the container inwhich the analytical reaction is to be run.

If volatile organic materials are present in the sample, plastic sample con-tainers cannot be used. Many small organic molecules are able to diffusethrough plastics. If the sample container is not tightly sealed, the volatiles may


escape into the environment. The loss will be greater if the sample is allowedto warm to room temperature. To avoid such losses, sample bottles must bemade of glass, with gas-tight seals. They should be filled to the top and storedat low temperature, but above freezing.

The most probable purely chemical reaction to be expected would be photo-lysis. This possibility can be prevented by storing the sample in the dark. Sincemost of these organic materials have been present together in seawater forsome considerable time, any other reactions are unlikely to occur without anincrease in temperature or the inadvertent addition of same catalyst. However,if biologically mediated reactions were to result in the formation of highlyreactive species, such as enzymes, further chemical reactions might then occur.Changes in pH, brought about either by biological activity or by the additionof mineral acids as preservatives, might also promote reactions.

Most of the changes taking place in stored seawater are due to biological,and principally bacterial, reactions. In one sense, it might seem strange thatremoving a sample of water from the ocean complete with its normal bacterialcomplement should result in any increase in bacterial activity. However, thefactor controlling bacterial activity seems to be the available surface area, sincefree flowing bacteria are not usually growing actively [48,49]. Thus, enclosinga sample of seawater in a bottle serves to furnish the free-floating bacteria witha large surface to which they can cling and upon which they can multiply. This“bottle effect” is well known, and must be prevented if the results of analysisare to have any meaning.

The two most popular methods of sample preservation are quick-freezingand the addition of inorganic poisons. Samples that are frozen without theaddition of preservatives are less likely to pick up contamination from thesample handling. If the sample bottles are to be frozen quickly enough toprevent bacterial growth, the sample bottles must be immersed in a freezingbath, to facilitate heat exchange. Such freezing baths are usually organic; thesealing and handling of sample bottles must be carried out with extreme care,in order to prevent contamination.

While it is normally considered that both biological and chemical reactionswill be essentially halted by freezing, this is not necessarily true. It has beenshown that some reactions of considerable biochemical importance are in factenhanced in the frozen state [50–52]. In any given case it cannot simply betaken for granted that freezing will be sufficient preservative; the efficiency ofthe methods must be tested for the compounds in question. The method is alsolimited to those analyses that can be performed on a small sample, perhaps100–200 ml. Larger volumes of seawater take too long to freeze. If the next stepin the sample preparation is to be freeze-drying, a considerable saving in time,as well as a decrease in possible contamination, can result from freezing thesample in the container to be used in the freeze-drying.

One is almost forced into hoping that freezing will prove satisfactory asa method of sample preservation, since none of the usual inorganic poisons


works ineverycase.Whilemercuric chloridehasbeen foundtobeeffectivewiththose marine organisms responsible for N2O production [53], it has been foundthat added radioactive glycine, in the presence of mercuric chloride in seawater,was decomposed slowly in the dark in the apparent absence of bacteria, theradioactive label appearing in thevapourphase.Of theother twomost favouredpoisons, sodium azide proved to be ineffective in stopping bacterial activityat concentrations as high as 10 mg/l; and cyanide, although effective at levelsbetween 1 and 10 mg/l, interfered with many of the chemical reactions.

Acidification with mineral acids is also often used as a method for preser-vation of organics. It would be expected that the distribution of organic com-pounds would be changed by such treatment, even if the total amount ofdissolved organic carbon were not appreciably altered. Certainly the volatilityof some of the organic molecules of low molecular weight is pH dependent;certain compounds normally retained at the slightly alkaline pH of seawaterwill be lost during freeze-drying or degassing of acidified samples.

Preservation of organic samples is thus still a major problem; there is nogeneral, foolproof method applicable to all samples and all methods of analysis.The most generally accepted method of sample preservation is storage underrefrigeration in the dark, with a preservative. This is another area that stillneeds extensive investigation.

Tables 1.7 and 1.8 show a selection of reagents used for preserving or fixingvarious determinands. These reagents are placed in the sample bottle before itis filled with sample; consequently the sample is “protected” from the momentit is taken.

The important influence that sample container materials can have on sea-water sample composition is illustrated next by two examples: one concerningthe storage of metal solutions in glass and plastic bottles, the other concerningthe storage of solutions of phthalic acid esters and polychlorinated biphenylsin glass and plastic.

1.3.1Losses of Silver, Arsenic, Cadmium, Selenium, and Zinc from Seawaterby Sorption on Various Container Surfaces [54]

The following container materials were studied: polyethylene, polytetrafluro-ethylene and borosilicate glass. The effect of varying the specific surfaceR (cm–1) was studied (ratio of inner container surface in contact with so-lution to volume of the solution) on adsorption of metals on the containersurface. New bottles were used exclusively. The differences in R values wereachieved by adding pieces of the material being considered. To avoid the possi-bility of highly active sites for sorption arising from fresh fractures, the edgesof the added pieces of borosilicate glass were sealed in a flame. Prior to use ofall materials the surfaces were cleaned by shaking with 8 M nitric acid for atleast three days and by washing five times with distilled water.


Tabl

e1.

7.D

etai

lsof

bott

les

and

fixin

gre

agen

ts(f

rom

auth

or’s

own

files

)

Ana

lysi

sB

ottl

em

ater

ial

Bot

tle

size

Rea

gent

sin

bott

leSh

elfl

ife

Not

es

Dis

solv

edox

ygen

Gla

ss12

5m

lN

one,

fixin

gki

tsup

plie

dU

nlim

ited

Toxi

cm

etal

sPo

lyet

hyle

ne25

0m

l1

ml

10m

l25%

Ana

lar

orA

rist

arH

NO

340

ml2

%H

NO

3

1M

onth

Ifso

lubl

em

etal

sto

bede

term

ined

,filte

rsam

ple

befo

reac

idad

diti

on.

Cle

anpl

asti

cC

asse

llaby

soak

ing

inac

id.

Phen

olG

lass

500

ml

1m

lH3P

O4

and

Ig

CuS

O4.

5H2O

Mad

eup

asne

eded

Silic

aPo

lyet

hyle

neN

one

Mad

eup

asne

eded

Mer

cury

Gla

ss12

5m

l7.

5m

l5%

K2C

r 2O

7pl

us4

mlc

onc.

H2S

O4

Bot

tles

fille

dw

ith

2N

HN

O3

whe

nno

tin

use;

rins

edou

tbe

fore

addi

ngre

agen

ts.

Chl

orop

hyll

Bro

wn

glas

s11

00g

Non

eU

nlim

ited

Tobe

kept

out

ofdi

rect

sunl

ight

afte

rfil

ling

wit

hsa

mpl

e.Su

lphi

deG

lass

(A)

110

gct

g(B

)R

eage

ntbo

ttle

ctg

5m

l132

gl–1

Zn

acet

ate

and

1m

lace

tic

acid

.96

gl–1

sodi

umca

rbon

ate

Tobo

ttle

alre

ady

cont

aini

ng5

ml

ofso

luti

on(A

)add

sam

ple

asqu

ickl

yas

poss

ible

,avo

idin

gen

trai

nmen

tofa

irbu

bble

s.5

mlo

fsol

utio

n(B

)is

then

adde

dto

fillt

hebo

ttle

,whi

chsh

ould

best

oppe

red

and

thor

ough

lym

ixed

.Pe

stic

ides

Gla

ss11

00g

Non

eU

nlim

ited

Bot

tle

chro

mic

acid

clea

ned,

grou

nd-g

lass

stop

pere

dan

dw

rapp

edin

poly

ethy

lene

bag.

Oils

Gla

ss11

00g

Non

eU

nlim

ited

Chr

omic

acid

clea

ned.

Sepa

rate

bott

lefr

omm

ain

sam

ple,

bott

le2/

3fil

led

wit

hsa

mpl

e.C

yani

deG

lass

diss

olve

dox

ygen

bott

les

2N

aOH

pelle

tsM

ade

upas

need

ed


Table 1.8. Sample preservation (from author’s own files)

Parameter Suggested methods Remarks

Acidity/Alkalinity

Refrigeration Gains or losses of CO2 affect the result. Microbialaction can affect this, should be one of the firstanalyses of the sample.

BOD Refrigeration Refrigeration is only partially effective. Analy-sis is best done immediately. Large changes canoccur over a few hours. Glass bottles preferred

COD 1 Refrigeration2 acidification with sulfuricacid to pH 2

Changes can occur quite rapidly with some sam-ples. Glass bottles preferred. DoE recommenda-tion. Hydrolysis can occur over longer periods.

Chlorine(residual)

Immediate analysis No storage possible.

Cyanide Addition of ascorbic acid,then render alkaline to pH11–12 with sodiumhydroxide

This determinand must be preserved immedi-ately. Need for separate container. Any residualchlorine must be removed with ascorbic acid.

Dissolvedoxygen

Winkler reagents Addition of reagent immediately after sampling.Fixed sample should be kept in the dark andanalysed as soon as possible. Where measure-ments are needed on liquids containing high lev-els of easily oxidisable organic matter, a portablemeter should be used, otherwise the copper sul-fate/sulfamic acid mix should be used. DoE rec-ommendation.

Metals(exceptingmercury)

DissolvedFilter on site into acid to pH1–2SuspendedFilter on siteTotalAcidify to pH 1–2 withhydrochloric or nitric acid

This procedure should be used wherever pos-sible at sewage works when compositing efflu-ents. DoE recommendations vary between 2 and20 ml, depending on the element. Polythene bot-tle required.For potable water, acidification must be carriedout on receipt in the laboratory; acidified bottlesshould not be left with sample takers.

Mercury Nitric acid (5 ml) + 5%potassium dichromate(5 ml) per litre

This treatment must be rendered immediatelyon sampling otherwise a large proportion canbe lost with in minutes. Glass bottle must beused (DoE recommendation).

Nitrogen(ammonia,nitrate,organic)

1 Sulfuric acid2 Chloroform3 Refrigeration

Chloroform as used effectively. Sulfuric acid iseffective and is frequently recommended. Acid-ification may be recommended by DoE.

Nitrite Chloroform Best analysed as soon as possible.pH Refrigeration Immediate analysis advisable, changes can oc-

cur quite rapidly, particularly if the container isopened. Well-buffered samples are less suscep-tible to changes.


Table 1.8. (continued)

Parameter Suggested methods Remarks

Phenolics Copper sulfate 1 g/l +phosphoric acid to reducepH to less than 4,refrigerated

Phosphate 1 Sulfuric acid conditioning2 Use of preparedbottle/refrigeration

Intensive investigation has been carried out onthis subject. The concentration of the determi-nand may affect the method. The DoE recom-mendation will be for low levels, i.e., less than25 µg/l, iodised plastic bottles for high levels,acid conditioned glass bottle.

Sulfide Zinc acetate/sodiumcarbonate, i.e., 5 mlzinc acetate (110 g zincacetate + 1 ml acetic acidper litre) + 5 ml sodiumcarbonate (80 g/l) to 100 mlsample

Must be preserved immediately. Care neededfor samples containing much suspended matter;may be DoE recommendation.

Surfactants,anionic

Refrigeration and either1 Mercuric chloride 1%(2 ml/l)2 Chloroform (2 ml/l)

These materials can be quite unstable over peri-ods longer than a few hours.

Surfactants,non-ionic

Formaldehyde (5 ml/l)

Working solutions (1 litre) which were 10–7 mol/l in one of the elements tobe studied were prepared by appropriate addition of the radioactive stock solu-tions to pH-adjusted artificial seawater. After the pH had been checked, 100 mlportions were transferred to the bottles to be tested. The filled bottles wereshaken continuously and gently in an upright position, at room temperatureand in the dark. At certain time intervals, ranging from 1 min to 28 d, 0.1 mlaliquots were taken. These aliquots were counted in a 3 × 3 in NaI (TI) well-type scintillation detector, coupled to a single-channel analyser with a windowsetting corresponding to the rays to be measured.

The counting times were chosen in such a way that at least 15 000 pulses werecounted. The sorption losses were calculated from the activities of the aliquotsand theactivityof thealiquot takenat timezero.Taking intoaccount thevarioussourcesof errors,mainly counting statistics, themaximumimprecision is about3%. Therefore, calculated sorption losses of 3% and lower are omitted fromthe listings as being not significant.

Table 1.9 shows the percentage loss as a function of time for silver, cadmium,and zinc from artificial seawater stored in polyethylene, borosilicate glass,PTFE at various pH and R values.


Tabl

e1.

9.So

rpti

onbe

havi

our

aspe

rcen

tage

ofsi

lver

,cad

miu

man

dzi

ncin

arti

ficia

lsea

wat

er[5

4]

Silv

erC

adm

ium

Zin

cM

ater

ial

Poly

ethy

lene

Bor

osili

cate

glas

sPT

FEB

oros

ilica

teB

oros

ilica

tegl

ass

glas

spH

8.5

48.

58.

54

48.

54

R(c

m–1

):1.

43.

41.

04.

21.

04.

21.

05.

51.

04.

21.

04.

21.

04.

21.

0co

ntac

tti

me

1m

in–

7–

––

––

–30

min

78

––

––

–3

––

––

––

–1

h6

5–

–3

3–

4–

––

–2

31–

2h

109

––

35

–6

––

––

910

1–

4h

1413

––

34

–7

––

––

929

58

h16

183

45

10–

8–

––

–9

304

24h

2428

44

69

612

––

––

528

52

d35

366

710

3113

17–

––

–4

264

3d

4445

611

3180

1323

5–

––

421

47

d64

6474

6027

7314

2914

010

43

185

14d

6672

8176

3984

2030

1343

2719

–9

421

d58

7780

7339

6426

3715

4125

173

105

28d

4678

8271

4067

2737

1436

2019

49

5

Cad

miu

min

poly

ethy

lene

and

inPT

FE,a

ndzi

ncin

poly

ethy

lene

,Rcm

–1=

spec

ific

surf

ace

area

.


For arsenic (added as sodium arsenate) and selenium (added as sodiumselenite), losses were insignificant in all the container materials considered,irrespective of matrix composition.

The sorption behaviour of trace elements depends on a variety of factorswhich, taken together,make sorption losses ratherdifficult topredict.However,the data from the study and from the literature indicate for which elementssorption losses may be expected as a function of a number of factors, such astrace element concentration, container material, pH, and salinity.

As was shown previously, reduction of contact time and specific surfacemay be helpful in lowering sorption losses, and acidification with strong acidwill generally prevent the problems of losses by sorption. However, it must beemphasised that the use of acids may drastically change the initial compositionof aqueous samples, making unambiguous interpretation of the analyticalresults cumbersome or even impossible [55].

For cases of sample storage where losses cannot be excluded a priori, somesort of check is required. This should be done under conditions representativeof the actual sampling, sample storage, and sample analysis. As this studyindicates, the use of radiotracers is helpful in making such checks.

The various factors in sorption losses may be classified into four cate-gories. The first category is concerned with the analyte itself, especially itschemical form and concentration. The second category includes the charac-teristics of the solution, such as the presence of acids (pH), dissolved material,(e.g., salinity, hardness), complexing agents, dissolved gases (especially oxy-gen, which may influence the oxidation state), suspended matter (competitorin the sorption process), and micro-organisms, (e.g., trace element take-up byalgae). The third category comprises the properties of the container, such asits chemical composition, surface roughness, surface cleanliness, and as thisstudy demonstrates, the specific surface. Cleaning by prolonged soaking in 8 Mnitric acid [56] is to be recommended. The history of the containers (e.g., age,method of cleaning, previous samples, exposure to heat) is important becauseit candirectly influence the type andnumberof active sites for sorption. Finally,the fourth category consists of external factors, such as temperature, contacttime, access of light, and the occurrence of agitation. All of these factors mustbe considered in assessing the likelihood of sorption losses during a completeanalysis.

Robertson[57]hasmeasured theadsorptionofzinc, caesium,strontium,an-timony, indium, iron, silver, copper, cobalt, rubidium, scandium, and uraniumonto glass and polyethylene containers. Radioactive forms of these elementswere added to samples of seawater, the samples were adjusted to the originalpH of 8.0, and aliquots were poured into polyethylene bottles, Pyrex-glass bot-tles and polyethylene bottles contained 1 ml concentrated hydrochloric acidto bring the pH to about 1.5. Adsorption on the containers was observed forstorage periods of up to 75 d with the use of a NaI(Tl) well crystal. Negligibleadsorption on all containers was registered for zinc, caesium, strontium, and


antimony. Losses of indium, iron, silver, copper, rubidium, scandium, and ura-nium occurred from water at pH 8.0 in polyethylene (excepting rubidium) orPyrex glass (excepting silver). With indium, iron, silver and cobalt, acidifica-tion to pH 1.5 eliminated adsorption on polyethylene but this was only partlyeffective with scandium and uranium.

Pellenberg and Church [58] have discussed the storage and processing ofestuarine water samples for analysis by atomic absorption spectrometry.

Flegal and Stokes [59] have described a sample processing technique neces-sary for avoiding lead contamination of seawater samples prior to lead stableisotope measurements by thermal ionisation mass spectrometry. Levels downto 0.02 ng/kg were determined.

Topreservemercury-containing samples [60,61]CoyneandCollins [61] rec-ommended preacidification of the sample bottle with concentrated nitric acidto yield a final pH of 1 in the sample solution. However, when this procedurewas used for storage of seawater, fresh water, and distilled deionised water inlow density polyethylene storage containers, abnormally high absorption wasobserved. This absorption at the mercury wavelength (253.7 nm) is due to thepresence of volatile organic plasticiser material and any polyethylene residueleached by the concentrated acidified nitric acid and by the acid solution atpH 1. These procedures employ sample storage in polyethylene bottles, andgas phase mercury detection may be subject to artificial mercury absorptiondue to the presence of organic material.

Reported mercury values in the oceans determined since 1971 span threeorders of magnitude, due at least in part to errors induced by incorrect sam-pling [62–64]. Olasfsson [65] has attempted to establish reliable data on mer-cury concentrations obtained in cruises in North Atlantic water.

The sampling, storage and analytical methods used by Olasfsson [65] inthis study have been evaluated. The Hydro-Bios water bottles used were mod-ified by replacing internal rubber rings with silicone rubber equivalents. Atthe commencement of a cruise the water bottles were cleaned by filling witha solution of the detergent Deacon 90. Samples for analysis of mercury weredrawn into 500 ml Pyrex vessels and acidified to pH 1 with nitric acid (Merck457), containing less than 0.05 nmol/l mercury impurities. The Pyrex bottleswere precleaned with both nitric acid and a solution of nitric and hydrofluoricacids (10:1), and subsequently stored up to the time of sampling holding a smallvolume of nitric acid. Ashore, reactive mercury was determined by cold vapouratomic absorption after preconcentration by amalgamation on gold [66]. Thetotal mercury concentration was similar to that obtained using a 500 W low-pressure mercury lamp (Hanovia) and immersion irradiation equipment. Theprecision of the mercury determination assessed by analysing 19 replicatesover a period of 107 d was found to be ±2.0 pmol/l for a concentration of12.5 pmol/l.


1.3.2Losses of Phthalic Acid Esters and Polychlorinated Biphenylsfrom Seawater Samples During Storage

During the storage of the sample, loss of analyte can occur via vaporisation,degradation, and/or adsorption. Adsorption of trace organic and inorganicspecies in seawater to container walls can severely affect the accuracy of theirdetermination. The adsorption of dichlorodiphenyltrichloroethylene [67] andhexachlorobiphenyl [68] onto glass containers has been observed.

Sullivan et al. [69] studied the loss of phthalic acid esters and chlorinatedbiphenyls from seawater whilst stored in glass containers. Equilibrium wasessentially reached in 12 h at 25 ◦C. Labelled compounds were used in some ofthe studies. Table 1.10 shows that between 2.2 and 49.9% of the organic soluteswere lost from the spiked solutions.

Table 1.10. Distribution of solutes among original solutions and subsequent water andsolvent rinses of containers [69]

Solute Initial aqueous con-centration (µg/l)

Percentage of solute recovered

Original spiked Water rinses of Acetonitrilewater test tubes rinses of test

tubes

DBP 4420 (140) 94.6 ND 5.485.4 ND 14.6

14C DBP 28.9 (2.2) 95.8 4.1 < 0.197.0 2.9 < 0.1

14C DBP 19.2 (2.0) 96.2 3.6 < 0.197.8 2.1 0.1

BEHP 407 (±9) 68.0 20.5 11.472.1 19.5 8.474.1 17.5 8.4

BEHP 229 (±7) 56.2 3.1 40.773.0 2.0 25.0

14C BEHP 15.5 (±0.3) 56.0 27.8 16.250.1 21.2 18.754.6 30.4 15.0

14C PCB (alone) 6.8 (±0.5) 71.1 5.5 23.481.1 0.6 18.3

14C PCB 6.8 (±0.5) 55.2 11.4 33.4(after BEHP)

55.3 4.4 40.3

DBP = Dibutyl phthalate; BEHP = Bis (2-ethylhexyl) phthalate; PCB = Polychlorinatedbiphenyls; ND = Not done

1.4 Sample Contamination During Analysis 27

The absorbed compounds were only partially recovered by subsequent wa-ter rinses. A solvent rinse of the containers recovered the remainder of thecompounds adsorbed to the surface.

The amount of phthalate bound to the glass test tubes appears to be a func-tion of the aqueous solubility of the phthalate. The solubilities of the phthalateshave been reported to be 3.2 mg dibutyl phthalate and 1.2 mg bis (2-ethylhexyl)phthalate per litre of artificial seawater [70]. Table 1.10 shows that the moresoluble dibutyl phthalate is absorbed far less than bis (2-ethylhexyl) phthalate.

The amount of solute also appears to be altered by the presence of other ma-terial already on the surface. Between 70% and 80% of the total polychlorinatedbiphenyls stayed in the aqueous phase when bis (2-ethylhexyl) phthalate wasnot present. When approximately 400 ng bis (2-ethylhexyl) phthalate was ab-sorbed to each test tube, only 55% of the total polychlorinated biphenyls stayedin the aqueous phase. The increased lipophilicity due to the presence of bis (2-ethylhexyl) phthalate apparently increased the adsorption of polychlorinatedbiphenyl by the glass test tubes.

The significant percentage of these organic pollutants in the unspiked waterrinses indicates that their adsorption is reversible. Once the compounds areadsorbed to the glass, they can desorb into the next solution that is placed intothe container, which is a concern during repeated use of the same storage orsampling container. The cross-contamination of samples and the loss of the an-alyte can be reduced by rinsing the container with clean water or solvent whichis then processed with the sample, or by extracting the sample in the container.

1.4Sample Contamination During Analysis

The environment in which samples are collected and processed during anoceanographic cruise would be considered unacceptable by any non-ocean-ographic microanalysis laboratory. On even the best-planned oceanographicvessels the space in which the samples are taken, the winch room and the wetlab, are normally awash in seawater, with a thin film of oil over most of theexposed surfaces. The worst case is to be found where the wet lab and winchroom are combined, or where the wet lab is the natural passageway betweenimportant parts of the ship, such as the engine room and the galley. Thesecircumstances are the rule rather than the exception on oceanographic vessels,even on those planned from scratch for oceanographic research. The reasonsfor these apparent flaws in planning are historical: chemical oceanographerswere interested either in major components of seawater or in trace nutrients,and neither of these kinds of analyses would be seriously damaged by thecontamination to be found in such wet labs.

With the recent emphasis upon the analysis of trace metals and organicmaterials, particularly possible pollutants, it has become obvious that cleanerworkingareasarenecessary.Thewinchroom,with its assortedgreases andoils,


must be separated from the sampling room, and ideally the people working onthe hydrographic wire, handling the samplers, should not also be drawing thesamples. The samplers should come into the sampling room through a hatchwhich can be closed. The sampling room should be a dead-end room, nota throughway, to discourage visitors. It would be unrealistic to expect a wetlab to be as free of contamination as a clean room, but it should approach theclean room in general arrangements. Even the air entering the sampling areashould be cleaned of hydrocarbons, perhaps by filtration through charcoal.The all-pervasive smell of diesel fuel in most oceanographic vessels does notbode well for the accuracy of any analyses for petroleum hydrocarbons.

If the analyses are to be performed on board, the room in which the samplesare prepared and analysed should, in fact, be built as a clean room. Manymodern oceanographic vessels are constructed to accept modular laboratories,which can be removed between voyages. Clean room modules, complete withair conditioning and filtered air supply, have been built for several vessels. It ispossible to perform accurate, precise microanalyses on board ship under lessfavourable conditions, but the analyst is really fighting the odds.

If the samples are to be brought back to a shore laboratory for analysis,contamination during analysis is more easily controlled. The analyst on shoremust have confidence in the people taking the samples: with the pressure onberth space and wire time on board ship, too often the samples are taken ona “while you’re out there, take some for me” basis. Again, ideally, the analystshould at least oversee every part of the process, from the cleaning of thesampler to the final calculation of the amounts present. His confidence in theaccuracy of the final calculation must decrease as he departs from the idealarrangement.

In the shore laboratory, the samples must be handled with the care neededfor any trace analysis. It must be remembered that the total amount of organiccarbon in seawater is around 1 ppm; single compounds are likely to be presentat ppb levels. In order to collect enough material even for positive identificationof some of the compounds present, the materials must often be concentrated.

Analytical chemists have long been aware of the necessity for purificationof any organic solvents used in trace analysis. The advent of gas and liquid–liquid chromatography has made plain just how many impurities can hidebehind a “high purity” label. Redistillation of organic solvents just before useis commonplace in most analytical laboratories. What has not been so evidentis the amount of organic material to be found in most inorganic reagents. Theactual amount present, let us say, in reagent grade sodium chloride may be lowenough so that it is not listed on the label, but still high enough to producean artificial seawater containing more organic carbon than the real thing.The presence of these compounds becomes serious when the analyst wishes toconcoct artificial seawater for standards and blanks. If the chemical in questioncan withstand oxidation, either at high temperature or in the presence of activeoxygen, the organic material may be eliminated. However, many compounds


used in routine analysis cannot be treated in this manner. If such chemicalsmust be used, the calculation of a true methods blank can become a majoranalytical problem.

Another problem, equally unrecognised, is the organic content of the dis-tilled water. For most analytical procedures, simple distillation is insufficienttreatment; perhaps in special cases, such extremes as distillation from per-manganate or distillation in quartz is considered necessary. For the analysisof organic materials at the ppm level in aqueous solutions, these methods arefar from sufficient. The experience of many workers has been that no form ofchemical pre-treatment will remove all of the organic material from distilledwater, and that some form of high-temperature oxidation of the impurities inthe water must be used [71–73]. Depending upon the original source of thewater, normal distillation will leave between 0.25 and 0.6 mg carbon/litre in thedistillate. The amounts and the kinds of compounds may vary with the seasonsandwith thedominantphytoplanktonspecies in the reservoirs. Sinceoceanwa-ter taken from depths greater than 500 m will usually contain only 0.3–0.7 mgcarbon/litre, it can be seen that the purity of the distilled water used to makeup reagents can be quite important. Sub-boiling distillation of deep seawatermight be an efficient starting point for the production of carbon-free blanks.

While at least partial solutions have been found to most of the problemsof contamination, these solutions have largely been adopted piecemeal by thevarious laboratories engaged in research on organic materials in seawater.The reasons for the adoption of half-way measures are largely historical. Thestudy of organic materials in seawater is relatively new, and the realisation thatdraconian measures are needed in the analysis is now becoming accepted.

The problems that can be encountered in sampling, sample preservation,and analysis have been discussed by King [74], Grice et al. [75] (sampling);Bridie et al. [76] (non-hydrocarbon interference); Farrington [77] (hydrocar-bon analysis); Kaplin and Poskrebysheva [78] (organic impurities); Hume [79],Riley [80] (oceanographic analysis); Acheson et al. [81] (polynuclear hydrocar-bons); Giam and Chan [82] (phthalates); Giam et al. [83] (organic pollutants);and Grasshoff [84, 85] (phosphates).

Mart [86] has described typical sample bottle cleaning routine for use whentaking samples for very low level metal determinations.

Sampling bottles and plastic bags, both made of high-pressure polyethylene,were rinsed by the following procedure. First clean with detergent in a labo-ratory washing machine, rinse with deionised water, soak in hot (about 60 ◦C)acid bath, beginning with 20% hydrochloric acid, reagent grade, followed bytwo further acid baths of lower concentrations, the latter being from Merck(Suprapur quality or equivalent). The bottles are then filled with dilute hy-drochloric acid (Merck, Suprapur), this operation being carried out on a cleanbench, rinsed and then filled with very pure water (pH 2). Bottles are wrappedinto two polyethylene bags. For transportation purposes, lots of ten bottles areenclosed hermetically in a large bag.


The determination of traces of heavy metals in natural waters can be greatlyaffected by contamination (positive and negative) during filtration and storageof samples [87–92]. Until the recent work of Scarponi et al. [93], this problem,especially as regards filtration, had not been studied adequately and system-atically with respect to the determination of cadmium, lead, and copper inseawater. Frequently, in order to make the determination easier, synthetic ma-trix samples [92–94] and/or high metal concentrations [92–96] have been used;otherwise, in order to demonstrate possible filter contamination, washed andunwashed filters have been analysed after washing [97–103].

The same filter can release or absorb trace metals depending on the metalconcentration level and the main constituents of the sample [92, 104], so theresults claimed must be considered with caution in working with naturalsamples. To avoid contamination, the following procedures have often beenused. Filters have been cleaned by soaking them in acids [88–90,100,103–107]or complexing agents [87, 88, 101] and/or conditioned either by soaking inseawater or a simulated seawater solution [10,104,108] or by passing a 0.2–2 lsample before aliquots are taken for analysis [88,90,103,109,110]. Sometimes,however, thewashingprocedurehasnotbeen found tobe fully satisfactory [87].For example, it has been reported that strong adsorption of cadmium andlead occurs on purified unconditioned membrane filters when triple-distilledwater is passed through the filter, whilst there is no change in the concentrationwith a river water sample after filtration of 500 ml [104]. Some investigatorsprefer to avoid filtration when the particulate matter does not interfere withthe determination (in which case the analysis must be completed soon aftersampling) [103, 111] or when open seawater is analysed. In the latter case,filtered and unfiltered samples do not seem to differ significantly in measurablemetal content [103, 112, 113].

The optimal conditions for uncontaminated long-term storage of diluteheavy metal solutions, particularly seawater, are now a topic of great interest.and contradictory results have frequently been reported. For example, withregard to the type of material to be used for the container, some workershave recommended the use of linear (high density) polyethylene instead ofconventional (low density) polyethylene [90, 114–116], while others have re-ported that linear polyethylene is totally unsuitable or inferior to low-densitypolyethylene [87, 89, 117–119]. Moreover, as a general rule, findings for par-ticular conditions are not necessarily applicable to elements, concentrations,matrices, containers, or experimental conditions different from those tested.As the macro and micro constituents of natural waters can differ widely [120],extreme caution must be used in handling published results. Possibly, as rec-ommended [88, 90, 121], it is best to ascertain the effectiveness of the storagesystem adopted in one’s own laboratory.

Scarponi et al. [93] used anodic stripping voltammetry to investigate thecontamination of seawater by cadmium, lead, and copper during filtration andstorage of samples collected near an industrial area. Filtration was carried


out under clean nitrogen to avoid sample contamination. Seawater leachesmetals from unclean membrane filters but, after one litre of water has passedthrough, the contamination becomes negligible. Samples stored in conven-tional polyethylene containers (properly cleaned and conditioned with pre-filtered seawater) at 4 ◦C and natural pH remain uncontaminated for threemonths (five months for cadmium); losses of lead and copper occur after fivemonths storage. Reproducibility (95% confidence interval) was 8–10%, 3–8%and 5–6% at concentration levels of about 0.06, 2.5, and 6.0 µg/l, for cadmium,lead, and copper, respectively.

The first aim of this work was to study the influence of an unwashed mem-brane filter on the cadmium, lead, and copper concentrations of filtered sea-water samples. It was also desirable to ascertain whether, after passage ofa reasonable quantity of water, the filter itself could be assumed to be cleanso that subsequent portions of filtrate would be uncontaminated. If this werethe case, it should be possible to eliminate the cleaning procedure and its con-tamination risks. The second purpose of the work was to test the possibilityof long-term storage of samples at their natural pH (about 8) at 4 ◦C, kept inlow-density polyethylene containers which have been cleaned with acid andconditioned with seawater.

Before use, new containers were cleaned by soaking in 2 M hydrochloric acidfor four days and conditioned with prefiltered seawater for a week, all at roomtemperature [90, 104]. Teflon-covered stirring bars (required for the voltam-metric measurements) were introduced into the containers at the beginning ofthe cleaning procedure. The containers used in another procedure and in thestudy of long-term storage could be regarded as having been conditioned forabout one month and more than two months, respectively. Other plastics usedin the sampling and filtration processes, and the components of the voltam-metric cell that came in contact with the sample solution, underwent the sameprocedure as the containers.

Figure 1.4 shows a typical curve demonstrating the dependence of concen-trations of copper, lead, and cadmium in the filtrate on the volume of seawatersampled. Metal levels become constant after 1–1.5 l of sample have been fil-tered, and it can be concluded that at this point, contamination of the sampleby the filtration equipment is negligible.

Table 1.11 gives the results of analytical measurements on aliquots of a con-ditioning seawater stock, stored at about 4 ◦C for three and five months inold low-density polyethylene containers (acid-washed for four days and con-ditioned for more than two months).

Apart from the observation that the concentrations are generally higherthan those measured previously, which indicates contamination during con-ditioning and manipulation, and the necessity of frequently renewing sea-water for equilibration purposes, it can be seen that there are no changes inthe metal concentrations for three months for lead and copper, or for fivemonths for cadmium. Also, after five months storage, some loss of lead and


Figure1.4. Concentration dependences on filtrate volume by procedure 1: (a) Cd; (b) Pb; (c)Cu. Numbers refer to storage time in days. • measured in order of sampling: ◦ measured inreverse order of sampling. From [93]

copper (21% and 24%, respectively), can be observed, possibly because ofthe formation and slow adsorption on container surfaces of hydroxo- andcarbonato-complexes [90, 122]. Hence at 4 ◦C in polyethylene containers nosignificant changes of heavy metal concentrations occur over a three-monthperiod [105, 110, 123].

Scarponi et al. [93] concluded that filtration of seawater through uncleanedmembrane filters shows positive contamination by cadmium, lead, and copper.In the first filtrate fractions, the trace metal concentration maybe increased bya factor of two or three. During filtration, the soluble impurities are leachedfrom the filter, which is progressively cleaned, and the metal concentrationin the filtrate, after passage of 0.8–1 l of seawater, reaches a stable minimumvalue. Thus it is recommended that at least one litre of seawater at naturalpH be passed through uncleaned filters before aliquots for analysis are taken


Table 1.11. Results after of long-term storage [93]

Date(1979)

Storagetime(months)

Metal concentration (µg/l)

Cadmium Lead CopperConc. Mean Change

(%)Conc. Mean Change

(%)Conc. Mean Change

(%)

July 2 0 0.160.19

0.17 4.0,4.3

4.2 7.7,8.0

8.0

0.170.16

4.3,4.2

8.4,8.0

Sept 30 3 0.150.16

0.16 – 6 5.0,3.6

4.1 – 2 7.8,7.5

7.8 – 2.5

0.160.18

4.2,3.6

8.1

Nov28th

5 0.180.18

0.17 0 3.7,3.1

3.3 – 21 5.9,5.8

6.05 – 24

0.16 2.6,3.8

6.5,6.0

from the subsequent filtrate. The same filter can be reused several times, andthen only the first 50–100 ml filtrate need be discarded. This system seemssimpler and more reliable for avoiding contamination than that of washingand conditioning filters before use, especially since in the latter case it hasbeen suggested that the first 0.2–2.0 l of filtrate should also be discarded.

Low-density polyethylene containers are suitable for storing seawater sam-ples at 4 ◦C and natural pH, provided that they are thoroughly cleaned (in2 M hydrochloric acid for at least a week) and adequately conditioned (withprefiltered seawater for at least one to two weeks). Storage can be prolongedfor at least three months (or five months for cadmium) without significantconcentration changes. For lead and copper, adsorption losses are observedafter five months.

The use of a special device that allows filtration under nitrogen, the directintroduction of sample into containers for storage during filtration and, theuse of these containers as analysis cells are all improvements that minimiseexternal sample contamination and improve between-sample reproducibility.

Degobbis [60] studied the storage of seawater samples for ammonia de-termination. The effects of freezing, filtration, addition of preservatives, andtype of container on the concentration of ammonium ions in samples storedfor up to a few weeks were investigated. Both rapid and slow freezing wereequally effective in stabilising ammonium ion concentration, and the additionof phenol as a preservative was effective in stabilising non-frozen samples forup to two weeks.


References

1. Ashton A, Chan R (1987) Analyst 112:8412. Schuessler U, Kremling K (1993) Deep-Sea Res Part 1 40:2573. Dunn J, Hall CD, Heath MR, Mitchell RB, Ritchie BJ (1993) Deep-Sea Res Part 1 40:8674. Teasdale P, Batley GE, Apte SC, Webster IT (1995) Trends Anal Chem 14:2505. Bufflap SE, Allen HE (1995) Water Res 29:20516. Yamamoto M, Urata K, Murashige K, Yamamoto Y (1981) Spectrochim Acta B 36:6717. Green DR (1977) Committee of Marine Analytical Chemistry, Canada Centre for Inland

Waters, Fisheries and Environment, Ottawa, Canada8. Wangersky PJ (1976) Ann Rev Ecol Sys 7:1619. Wangersky PJ (1976) Deep-Sea Res 23:457

10. Gordon Jr., DC, Keizer PD (1974) Technical Report No. 481 Fisheries Research Board,Canada

11. Harvey GW (1966) Limnol Oceanogr 11:60812. Garrett WD (1965) Limnol Oceanogr 10:60213. Daumas RA, Laborde PL, Marty JC, Saliot A (1976) Limnol Oceanogr 319:31914. Harvey GW, Burzell LA (1972) Limnol Oceanogr 19:15615. Garrett WD, Burger R (1974) Sampling and determining the concentration of film-

forming organic constituents of the air–water interface, US NRL Memo Report 285216. Larsson K, Odham G, Sodergren A (1974) Mar Chem 2:4917. Miget R, Kator H, Oppenheimer C (1974) Anal Chem 45:115418. Anufrieva NM, Gornitsky AB, Nesterova MP, Neirovskaya IA (1976) Okeanologiya

16:25519. Baier RE (1972) J Geophys Res 77:506220. Gump BH, Hertz HA, May WE, Chesler SN, Dyszel SM, Enagnio DP (1975) Anal. Chem

47:122321. Keizer PD, Gordon DC, Jr., Dale J (1977) Journal of the Fisheries Research Board 34:3722. Clark Jr., RC, Blumer M, Raymond SO (1967) Deep-Sea Res 14:12523. Bertoni R, Melchiorri-Santolini U (1972) Mem Inst Ital Hydrobiol 29:9724. Gordon Jr., DC, Keizer PD, Dale J (1974) Mar Chem 251:25125. Smith KL (1971) Limnol Oceanogr 16:67526. Bewers JM, Dalziel J, Yeats PA, Barron JL (1981) Mar Chem 10:17327. Olaffson J (1978) Mar Chem 87:8728. Olaffson J (1980) A preliminary report on ICES intercalibration of mercury in seawater

for the joint monitoring group of the Oslo and Paris commissions, submitted to theMarine Chemistry Working Group of OCS, February 1980

29. Thibaud Y (1980) Exercise d’Intercalibration CIEM, 1979, Cadmium en eau de mer,report submitted to the Marine Chemistry Working Group of ICES

30. Jones PGW (1977) A preliminary report on the ICES intercalibration of sea watersamples for the analyses of trace metals, ICES CM1977/E:16

31. Jones PGW, (1976) An ICES intercalibration exercise for trace metal standard solutions,ICES CM1979/E:15

32. Windom HL (1979) Report on the Results of the ICES Questionnaire on Sampling andAnalysis of Water for Trace Elements, Submitted to the First Meeting of the MarineChemistry Working Group, Lisbon

33. Pocklington R (1972) Variability of Ocean off Bermuda, Bedford Institute of Oceanog-raphy Report Series BI-R-72-73

34. Boyle EA, Sclater FR, Edmond JM (1977) Sci Lett 37:3835. Bruland KW, Knauer GA, Martin JH (1978) Limnol Oceanogr 23:618

References 35

36. Bruland KW, Knauer GA, Martin JH (1978) Nature 271:74137. Sclater FR, Boyle EA, Edmond JM (1976) Earth Planet Sci Lett 31:11938. Wong CS, Kremling K, Riley JP et al. (1985) Ocean Chemistry Division Institute of

Ocean Sciences, PO Box 6000, Sidney, BC, V8L 4B2, Canada. Ocean Chemistry Divisioncontract to SEAKEM Oceanography Ltd., Sidney, BC, Canada. Marine Chemistry De-partment, Institut für Meereskunde und der Universität, Kiel, Dunsternbrooker Weg20, 2300 Kiel, FR Germany. Department of Oceanography, University of Liverpool, POBox 147, Liverpool, L69 3BX, UK

39. Patterson CC, Settle M (1976) The reduction of orders of magnitude errors in leadanalysis. In: LaFleur PD (ed) Accuracy in trace analysis: sampling, sample handling,analysis. NBS Special Publication 422, p. 321

40. Bothner MH, Robertson DE (1975) Anal Chem 47:59241. Participants in IDOE Workshop (1976) Mar Chem 4:38942. Stukas VJ, Wong CD (1976) Science 211:142443. Wong CS, Kremling K, Riley JP et al. (1979) Accurate Measurement of Trace Metals in

Sea Water: an Intercomparison of Sampling Devices and Analytical Techniques usingCEPEX Enclosure of Sea Water. Unpublished manuscript report, NATO study fundedby NATO Scientific Affairs Division

44. Abdullah MI, El-Rayis OA, Riley JP (1976) Anal Chim Acta 84:36345. Abdullah MI, Royale LG (1972) Anal Chim Acta 80:5846. Danielsson LG, Magnusson B, Westerlund S (1978) Anal Chim Acta 98:4747. Otsuki A, Hanya T (1972) Limnol Oceanogr 87:24848. Jannasch NW, Pritchard PH (1972) Mem Inst Ital Hydrobiol 24 Suppl:28949. Wiebe WS, Pomeroy LR (1972) Mem Inst Ital Hydrobiol 29 Suppl:32550. Alburn HE, Grant NH (1965) J Am Chem Soc 87:417451. Grant NH, Alburn HE (1965) Biochemistry 4:191352. Grant NH, Alburn HE (1967) Arch Biochem Biophys 118:29253. Yoshinari I (1976) Mar Chem 4:18954. Massee R, Maessen FJMJ (1981) Anal Chim Acta 127:18155. Florence TM, Bailey GM (1980) CRC Critical Reviews on Analytical Chemistry Au-

gust:21956. Karin RW, Buone JA, Fashing JL (1975) Anal Chem 47:229657. Robertson DE (1968) Anal Chim Acta 42:53358. Pellenburg RE, Church TM (1978) Anal Chim Acta 97:8159. Flegal A, Stukas V (1987) J Mar Chem 22:16360. Degobbis D (1973) Limnol Oceanogr 18:14661. Coyne R, Collins JA (1972) Anal Chem 44:109362. Carr RA, Wilkniss PE (1973) Environ Sci Technol 7:6263. Feldman C (1974) Anal Chem 46:9964. Fitzgerald WF (1979) Distribution of mercury in natural waters. In: Nriagu JO (ed)

The Biogeochemistry of Mercury in the Environment. Elsevier, Amsterdam, The Nether-lands

65. Olaffson J (1981) Trace Metals in Seawater. In: Wong CS et al. (eds) Proceedings ofa NATO Advanced Research Institute on Trace Metals in Seawater, 30/3-3/4/81, Sicily,Italy. Plenum Press, New York, NY, USA

66. Olaffson J (1974) Anal Chim Acta 68:20767. Picer M, Picer N, Strohal P (1977) Sci Total Environ 8:15968. Pepe HG, Byrne JJ (1980) Bull Environ Contam Toxicol 25:93669. Sullivan KR, Altas EL, Giam CS (1981) Anal Chem 53:71870. Kakreka JPMS (1974) Thesis, Texas A&M University, USA


71. Wangersky PJ (1965) American Science 53:35872. Hickman K, White I, Stark E (1973) Science 180:1573. Conway BE, Angerstein-Kozlowska H, Sharp WBA (1973) Anal Chem 45:133174. King DL, Ciaccio LL (1971) (eds) Sampling of Natural Waters and Waste Effluents,

Marcel Dekker, New York, NY, USA, p. 45175. Grice GD, Harvey GR, Bown VT, Backus RH (1972) Bull Environ Contam Toxicol 7:12576. Bridie AL, Box J, Herzberg S (1973) J Inst Petrol 59:26377. Farrington JW (1974) AD Report No. 777695/GA, US National Technical Information

Service78. Kaplin AA, Poskrebysheva LM (1974) Izuestia Tomak Politckh Inst 233:9179. Hume DN (1975) Fundamental problems in oceanographic analysis. In: Gibb, Jr., RP

(ed) Analytical Methods in Oceanography. ACS, Washington, DC, USA, p. 180. Riley JP (1975). In: Riley JP, Skirrow G (eds) Chemical Oceanography, Volume 3.

Academic Press, London, UK, p. 19381. Acheson MA, Harrison RM, Perry R, Wellings RA (1976) Water Res 10:20782. Giam CS, Chan HS (1976) Special Publication, National Bureau Standards (US), No.

422, p. 76183. Giam CS, Chan HS, Neff GS (1976) In: Proceedings of International Conference Envi-

ronmental Sensing Assessment ACS, Washington, DC, USA84. Grasshoff K (1966) Anal Chem 220:8985. Grasshoff K (1976) Verlag Chemie 1:5086. Mart L (1979) Fresen Z Anal Chem 296:35087. Robertson DE (1972) In: Ultrapurity Methods and Techniques, Zeif M, Speights R (eds).

Marcel Dekker, New York, NY, USA, p. 20788. Riley JP, Robertson DE, Dutton JWR et al. (1975). In: Riley JP, Skirrow G (eds) Chemical

Oceanography, 2nd Edition, Volume 3. Academic Press, London, UK, p. 19389. Zief M, Mitchell JW (1976) Contamination Control in Trace Element Analysis, Wiley

New York, NY, USA90. Batley GE, Gardner D (1977) Water Res 11:74591. Salim R, Cooksey G (1980) J Electroanal Chem 106:25192. Truitt RE, Weber JH (1979) Anal Chem 51:205793. Scarponi G, Capodaglio G, Cescon P, Cosma B, Frache R (1982) Anal Chim Acta 135:26394. Weber JH, Truitt RE (1979) Research Report 21, Water Resource Research Center,

University of New Hampshire, Durham, NH, USA95. Marvin KT, Proctor, Jr., RR, Neal RA (1970) Limnol Oceanogr 15:32096. Gardiner J (1974) Water Res 8:15797. Spencer DW, Manheim FT (1969) US Geological Survey Professional Paper, 650-D, p.

28898. Spencer DW, Brewer P, Sachs PL (1972) Geochim Cosmochim Acta 36:7199. Dams R, Rahn KA, Winchester JW (1972) Environ Sci Technol 6:441

100. Wallace, Jr., GT, Fletcher IS, Duce RA (1977) J Environ Sci Health A12:493101. Duychserts G, Gillain G (1977). In: (ed Wanninen W) Essays on Analytical Chemistry

Pergamon, Oxford, UK, p. 417102. Smith RG (1978) Talanta 25:173103. Mart L (1979) Fresen Z Anal Chem 296:352104. Nurnberg HW, Valenta P, Mart L, Raspor B, Sipos L (1976) Fresen Z Anal Chem 282:357105. Batley GE, Gardner D (1978) Estuar Coast Mar Sci 7:59106. Bruland KW, Franks RP, Knauer GA (1979) Anal Chim Acta 105:233107. Burrell DC (1979) Mar Sci Comm 5:283108. Figura P, McDuffie B (1980) Anal Chem 52:1433

References 37

109. Burrell DC, Lee ML (1975) Water Quality Parameters, ASTM STP-573, ASTM, p. 58110. Fukai R, Huynh-Ngoc L (1976) Mar Pollut Bull 7:9111. DeForest A, Pettis RW, Fabris G (1978) Australian Journal of Marine and Freshwater

Research 29:193112. Fukai R, Huynh-Ngoc, L (1976) Anal Chim Acta 83:375113. Zirino A, Lieberman SH, Clavell C (1978) Environ Sci Technol 12:73114. Gardiner J, Stiff MJ (1975) Water Res 9:517115. Bowen VT, Strohal P, Saiki M et al. (1970) In: Reference Methods for Marine Radioactiv-

ity Studies (eds) Nishiwaki Y, Fukai R, International Atomic Energy Agency, Vienna,Austria, p. 12

116. Bowditch DC, Edmond CR, Dunstan PJ, McGlynn J (1976) Technical Paper No. 16,Australian Water Resources Council, p. 22

117. Tolg G (1972) Talanta 19:1489118. Tolg G (1975). In: (ed) Svehia G. Comprehensive Analytical Chemistry, Volume 3 Else-

vier, Amsterdam, The Netherlands119. Moody JR, Lindstrom RM (1977) Anal Chem 49:2264120. Davison W, Whitfield M (1977) J Electroanal Chem 75:763121. Scarponi G, Miccoli E, Frache R (1978). In: Proceedings of the 3rd Congress of the

Association of Italian Oceanology and Limnology, Sorrento, Italy. Pergamon Press,Oxford, UK, p. 433

122. Subramanian KS, Chakrabati CL, Sheiras JE, Maines IS (1978) Anal Chem 50:444123. Carpenter JH, Bradford WL, Grant V (1975) In: (ed) Cronin LE. Estuarine Research

Academic Press, New York, NY, USA, p. 188

2 Determination of Anions

2.1Acetate

2.1.1Ion Chromatography

Xiao-Hua Yang et al. [1] determined nanomolar concentrations of individuallow molecular weight carboxylic acids (and amines) in seawater. Diffusion ofthe acids across a hydrophobic membrane was used to concentrate and separatecarboxylic acids from inorganic salts and most other organic compounds priorto the application of ion chromatography. Acetic propionic acid, butyric-1 acid,butyric-2 acid, valeric and pyruvic acid, acrylic acid and benzoic acid were allfound in reasonable concentrations in seawater.

2.2Acrylate


See Sect. 2.1.1.

2.3Alkalinity

2.3.1Titration Method

Among the possible analytical methods for alkalinity determination, Gran-type potentiometric titration [2] combined with a curve-fitting algorithm isconsidered a suitable method in seawaters because it does not require a prioriknowledge of thermodynamic parameters such as activity coefficients anddissociation constants, which must be known when other analytical methodsfor alkalinity determination are applied [3–6].

40 2 Determination of Anions

Alkalinity determination in hypersaline solutions by Gran-type titration issubject to a number of errors which can usually be neglected in lower ionicstrength solutions. For example, the pH readings along the titration path maybe inaccurate due to the marked difference between the ionic strength andcomposition of the sample and the standard buffers used to calibrate the glassand reference electrode pair [7].

Ben Yaakov and Lorch [8] identified the possible error sources encoun-tered during an alkalinity determination in brines by a Gran-type titrationand determined the possible effects of these errors on the accuracy of the mea-sured alkalinity. Special attention was paid to errors due to possible non-idealbehaviour of the glass-reference electrode pair in brine. The conclusions ofthe theoretical error analysis were then used to develop a titration procedureand an associated algorithm which may simplify alkalinity determination inhighly saline solutions by overcoming problems due to non-ideal behaviourand instability of commercial pH electrodes.

The titration procedure used was that described by Ben Yaakov et al. [9], inwhich 100 ml of seawater or brine sample is titrated in 30–50 burette incre-ments with 3–5 ml standard 0.1 M hydrochloric acid using a pH electrode.

The accuracy of the method was tested experimentally by running dupli-cate titrations on distilled water, artificial seawater with and without sulfate,and artificial Dead Sea waters. For each run, alkalinity was calculated by twomethods:

1. by the conventional Gran plot, which presumes that the glass electrode isproperly calibrated;

2. by the method that applies the titration data for in situ calibration of theglass electrode via the slope correction algorithm.

The precision of the latter method when applied in the distilled water runs wasfound to be significantly better than the conventional method. This could beattributed to the non-stability of the glass electrode, which is corrected for bythe proposed algorithm.

2.3.2Spectrophotometric Methods

Various workers have discussed the determination of total alkalinity and car-bonate [10–12], and the carbonate:bicarbonate ratio [12] in seawater. A typicalmethod utilises an autoanalyser. Total alkalinity (T: milliequivelents per litre)is found by adding a known (excess) amount of hydrochloric acid and backtitrating with sodium hydroxide solution; a pH meter records directly and afterdifferentiation is used to indicate the end-point. Total carbon dioxide (C: mil-liequivelents per litre of HCO–

3 per litre) is determined by mixing the samplewith dilute sulfuric acid and segmenting it with carbon dioxide-free air, sothat the carbon dioxide in the sample is expelled into the air segments. The air

2.4 Arsenate/Arsenite 41

is then separated from the sample and passed into buffered phenolphthaleinsolution, thereby lowering the pH and diminishing the colour of the phenolph-thalein. The reduction in colour is measured colorimetrically (540 nm). Theconcentration of carbonate is given by 2(T–C) milliequivelents per litre, andthe concentration of bicarbonate is 2C–T milliequivelents per litre.

A computer program has been used to calculate the magnitude of systematicerrors incurred in the evaluation of equivalence points in hydrochloric acidtitrations of total alkalinity and carbonate in seawater by means of Gran plots.Hansson [13] devised a modification of the Gran procedure that gives improvedaccuracy and precision. The procedure requires approximate knowledge of allstability constants in the titration.

2.4Arsenate/Arsenite

2.4.1Spectrophotometric Method

Haywood and Riley [14] have described a spectrophotometric method for thedetermination of arsenic in seawater. Adsorption colloid flotation has beenemployed to separate phosphate and arsenate from seawater [15]. These twoanions, in 500 ml filtered seawater, are brought to the surface in less than 5 min,byuseof ferrichydroxide (addedas0.1 MFeCl2; 2 ml) as collector, atpH4, in thepresence of sodium dodecyl sulfate [added as 0.05% ethanolic solution (4 ml)]and a stream of nitrogen (15 ml/minutes). The foam is then removed and phos-phate and arsenate are determined spectrophotometrically [16]. Recoveries ofarsenate and arsenite exceeding 90% were obtained by this procedure.

2.5Benzoate


See Sect. 2.1.1.

2.6Butyrate


See Sect. 2.1.1.


2.7Borate


Sato [17] has examined several α-hydroxy acids as complexing agents for theextraction-spectophotometric determination of boron. Mandelic acid is themost useful. The boron-complex anion obtained is extracted into benzenewith malachite green in a single extraction; boron is determined indirectly bymeasuring the absorbance of malachite green in the extract at 633 nm. Thecalibration graph is linear over the range 7.50 ×10–7 –1.50 ×10–5 mol/l boron;the apparent molar absorptivity is 6.52 ×104 l/mol/cm. The method is appliedto the determination of micro amounts of boron in waters with satisfactoryresults.

2.8Bromate

2.8.1Spectrophotometric Titration and Differential Pulse Polarography

Chlorinated waters are being discharged to estuaries and coastal waters inincreasing quantities. In such systems the chlorine reacts with the naturalbromide and ammonia at pH 8 to produce the highly toxic hypobromousacid, hypobromite ion, and haloamines. For normal seawater of pH 8, theinitial products of chlorination are a mixture of hypobromous acid and thehypobromite ion. Both of these compounds are unstable with respect to de-composition and disproportionation.

Macalady et al. [18] report experiments in which chlorinated sea water wasexposed to sunlight at various intensities, to simulate different periods of theday. The results of subsequent analyses for bromates and residual oxidantsshow that the rate and extent of bromate formation depend on the intensityof the sunlight, with none found in samples kept in the dark for 24 h at 40 ◦C.It would appear that large amounts of bromate have already been producedin estuarine and coastal waters with unknown effects, as little information isavailable on the direct toxicity of the bromate ion.

Details of the chlorination experiment referred to previously are now dis-cussed in more detail.

After chlorination, beakers were removed from the sunlight at regular (usu-ally 30-min) intervals, placed in a dark box, and the contents analysed forbromate and residual oxidants without delay. Residual oxidants analyses wereperformed by the I–

3 spectrophotometric titration procedure described by Car-penter [19] with a pH of 2 and potassium iodide concentration of 4 g/l. Bromate

2.8 Bromate 43

analyses were made by differential pulse polarography at 25 ◦C and a pH of8.35 (after oxygen stripping with nitrogen), using a Princeton Applied Researchmodel 174A polarographic analyser.

A typical polarographic recording is shown in Fig. 2.1; curve (a) is the po-larogram obtained for chlorinated seawater analysed immediately after chlo-rination. Identical traces were observed for non-chlorinated seawater and forchlorinated seawater kept in the dark for periods up to 24 h at temperatures upto 40 ◦C, which indicates a lack of bromate formation under these conditions(BrO–

3 ≤ 10–7 M, less than 0.5% conversion of chlorine). Addition of coppersulfate to give a cupric ion concentration in the seawater of 100 parts per billiondid not induce measurable bromate production in the dark. Curve (b) was ob-tained from a chlorinated (4.9 mg/l) seawater solution that was exposed to fullsunlight for 70 min. Curve (c), which is offset by 0.4 µA with respect to curves(a) and (b), shows the presence of 1.0 ×10–5 M sodium bromate in seawater.

Figure 2.2 illustrates kinetic data for the appearance of bromate (Fig. 2.2a)and disappearance of residual oxidants (Fig. 2.2b) in chlorinated seawater ex-

Figure 2.1. Differential pulse polarographic verification of sunlight-induced bromate pro-duction in chlorinated seawater. Curve (a) polarogram from untreated seawater, seawaterimmediately after chlorination to 4.9 ppm, or chlorinated seawater kept in the dark for4 h at 40 ◦C. Curve (b) polarogram from chlorinated seawater exposed to full sunlight for70 min. Curve (c) standard: 1.0 ×10–6 M sodium bromate in seawater, offset with respect tocurves (a) and (b). Polarograms were recorded at 25 ◦C and pH 8.35; SCE: saturated calomelelectrode. From [19]


Figure2.2. (a)Disappearancewith timeof residualoxidantsand(b) concomitantappearanceof bromate in chlorinated seawater (4.2–4.9 ppm of chlorine) as a function of exposure tosunlight. The conditions were: Curve (a) full midday sunlight, Curve (b) 65% of full sunlight,and Curve (c) overcast, 20% of full sunlight. Curve (d) shows residual oxidant disappearancein the dark at 40 ◦C. No bromate production was observed in the dark. From [19]

posed to sunlight. Curves were obtained from solutions exposed to full middaysunlight for the duration of the experiment: curves (b) were for exposure topartial sunlight (the average intensity was approximately 65% of full sunlight);and curves (c) were for overcast conditions (average light intensity, 20% of fullsunlight). Curve (d) in Fig. 2.2a shows the disappearance of residual oxidantswith time at 40 ◦C in the dark. The ordinates are calibrated as the percentageof the added chlorine recovered as residual oxidants (Fig. 2.2a) or as bromateformed by decomposition of hypobromite.

The lack of observable bromate production in the dark is not inconsis-tent with the report of Lewin and Avrahani [20] that substantial bromate wasformed in their 0.05 M hypobromide solutions. The solutions used by Macalady

2.9 Bromide 45

et al. [18], which correspond to current chlorine use in water chlorination pro-cesses, were 1000 times more dilute. Using their rate constants, they calculatedin their solutions a conversion to bromate of less than 1% after 24 h.

The loss of residual oxidants does not correspond exclusively to bromateformation, and other reactions, including oxidation of organic matter, alsotake place. The rate and extent of bromate formation depend on the intensityof sunlight.

Solutions of 10–5 sodium bromate in seawater were also titrated by using theresidual oxidants procedure. The results show that only about 5% of the addedbromate appears as residual oxidants. Thus, the apparent residual oxidantsremaining after exposure to sunlight could be completely due to bromate.

2.9Bromide


Bromide in seawater can be determined by the procedure described below,which is capable of determining content down to 0.1 mg/l bromide.

The sample is acidified with sulfuric acid. The bromide content is then de-termined by the volumetric procedure described by Kolthoff and Yutzy [21]. Inthis procedure the buffered sample is treated with excess sodium hypochloriteto oxidise bromide to bromate. Excess hypochlorite is then destroyed by ad-dition of sodium formate. Acidification of the test solution with sulfuric acidfollowed by addition of excess potassium iodide liberates an amount of iodineequivalent to the bromate (i.e., the original bromide) content of the sample.The liberated iodine is titrated with standard sodium thiosulfate.

2.9.2X-ray Emission Spectrometry

Rose and Cuttita [22] proposed a graphical method for evaluating the back-ground when determining traces of bromide by X-ray emission spectrography(carried out on cellulose pellets containing the residue obtained by evapora-tion of the sample in the presence of standard bromide solution). Based onthis work a graphical method is proposed which provides a value (mB) of theamount of bromide equivalent to the background, which is substituted intothe equation: counts per sec = k(m + ms + mB), where k is a factor thatdepends on the sample, mB is the bromide content of the original solution,and ms is the amount of bromide added as “spike”. Results obtained by thetwo methods agree satisfactorily, but it is stressed that both are approxima-tions.


2.9.3Segmented Flow Analysis

Basel et al. [23] have described methods of compensating for chloride, am-monia, and bicarbonate interferences in determining bromide in saline waterswith an automated segmented flow analyser utilising the phenol red method.

2.9.4Solid State Membrane Electrodes

Walters [24] examined the effect of chloride on the use of bromide and iodidesolid state membrane electrodes, and he calculated selectivity constants. Mul-tiple linear regression analysis was used to determine the concentrations ofbromide, fluorine, and iodide in geothermal brines, and indicated high inter-ferences at high salt concentrations. The standard curve method was preferredto the multiple standard addition method because of:

1. the deviation from linearity at high salt concentrations of the bromideelectrode;

2. the loss of accuracy due to increase in sample volume by using volumeincrements;

3. the limitations in reading from the Orion meter (0.1 mV);4. the fact that bromide, fluoride, and iodide were present at relatively low

concentrations (where the electrode exhibited nonlinear behaviour).

2.9.5X-ray Fluorescence Spectroscopy

Sichère et al. [25] determined bromine concentrations in the 0.06–120 mg/lrange in brines, directly by X-ray fluorescence using selenium as an in-ternal standard to eliminate interference effects. Lower concentrations ofbromine must be concentrated on filter paper containing an ion exchangeresin. The same concentrations of chlorine can be determined with the ad-dition of barium to reduce the interferences from carbonates and sulfates.Relative standard deviation was better than 1%. The interference of someother ions (e.g., calcium, potassium, magnesium, sodium, and iron) was ex-amined.

2.9.6Isotachoelectrophoresis

Fukushi and Hiro [26] have described a capillary-type isotachoelectrophoreticmethod for the determination of bromide in seawater.

2.10 Chloride 47

2.10Chloride


Chlorine at the percentage level at which it occurs in sea water is usuallydetermined by classical procedures using standard silver nitrate as the titrantand potassium chromate indicator, or alternatively by the mercuric thiocyanateprocedure using dithizone as indicator. As large dilutions of the original sampleare involved in these analyses, it is essential to use grade A glassware and takeall other suitable precautions, such as temperature control.

Chloride can also be estimated by potentiometric titration using standardsilver nitrate [27]. The results are recorded directly and evaluated by means ofa computer program based on the Gran extrapolation method. The determi-nations have a precision of ±0.02% and since many samples can be titratedsimultaneously, the time for a single determination can be reduced to less than5 min.

Jagner [28] has also described a semi-automatic titration for high-precisiondetermination of chlorine in seawater, where it has been used for the poten-tiometric determination of total halides (silver electrode) and alkalinity (glasselectrode), and for the photometric titration of total alkaline-earth metals.Several titrations can be effected simultaneously.

Grasshoff and Wenck [29] have described a version of the Mohr–Knudsensilver nitrate procedure for the determination of the chlorinity of sea water. Inthis method, which overcomes the disadvantages of conventional burettes, useis made of a motor-driven piston burette of 20 ml capacity, which is sufficientfor a chlorinity range of 0–45 per thousand. The accuracy is the same as forconventional titration. The apparatus is compact and portable.

Several autoanalyser procedures are available for the determination of chlo-rine.

2.10.2Ion Selective Electrodes

Noborn[30]has studied thedynamicpropertiesof chloride selective electrodesand their application to sea water. An Orion 94-17 electrode was used in thesestudies.

In the presence of bromide ions the electrode was subject to a drop inpotential, (e.g., 1.5 to 5.7 mV at a Br–:Cl– ratio of 2000:3) and to delayedresponse. A considerable hysteresis effect is also observed in concentratedsolutions of chloride when the electrode is used in a 1 M chloride solution andthen dipped in one that is 0.02 M. Equilibrium is reached only after 10 min.The junction potential is minimised by diluting the test solution with thesalt-bridge solution (10% aq. potassium nitrate).


2.10.3Chronopotentiometry

Chronopotentiometry has also been used to determine chloride ions in sea-water [31]. The chloride in the solution containing an inert electrolyte was de-posited on a silver electrode (1.1 cm2) by the passage of an anodic current. Thecell comprised a silver disc as working electrode, a symmetrical platinum-disccounter-electrode and a Ag–AgCl reference electrode to monitor the potentialof the working electrode. This potential was displayed on one channel of a two-channel recorder, and its derivative was displayed on the other channel. Thechronopotentiometric constant was determined over the chloride concentra-tion range 0.5 to 10 mM, and the concentration of the unknown solution wasdetermined by altering the value of the impressed current until the observedtransition time was about equal to that used for the standard solution.

2.10.4Miscellaneous

Wilson [32] has described a portable flow-cell membrane salinometer. Al-though test solutions are normally passed through the cell, the electrodes canbe connected to a remote membrane sensor head by means of salt bridges formeasurement of the salinity of estuarine muds in situ. The error is within ±1%over the salinity range 1–40%.

2.11Chromate and Dichromate

2.11.1Atomic Absorption Spectrometry

A method has been developed for differentiating hexavalent from trivalentchromium [33]. The metal is electrodeposited with mercury on pyrolyticgraphite-coated tubular furnaces in the temperature range 1000–3000 ◦C, us-ing a flow-through assembly. Both the hexa- and trivalent forms are depositedas the metal at pH 4.7 and a potential at –1.8 V against the standard calomelelectrode, while at pH 4.7, but at –0.3 V, the hexavalent form is selectively re-duced to the trivalent form and accumulated by adsorption. This method wasapplied to the analysis of chromium species in samples of different salinity, inconjunction with atomic absorption spectrophotometry. The limit of detectionwas 0.05 µg/l chromium and relative standard deviation from replicate mea-surements of 0.4 µg chromium (VI) was 13%. Matrix interference was largelyovercome in this procedure.

Various workers have discussed the separate determination of Cr(III) andCr(VI) in seawater [34–38].

2.11 Chromate and Dichromate 49

Cranston and Murray [35,36] took samples in polyethylene bottles that hadbeen pre-cleaned at 20 ◦C for four days with 1% distilled hydrochloric acid.Total chromium Cr(VI) + Cr(III) + Crp (Crp: particulate chromium) was co-precipitated with iron (II) hydroxide, and reduced chromium Cr(III) + Crpwas co-precipitated with iron (III) hydroxide. These co-precipitation stepswere completed within minutes of the sample collection to minimise storageproblems. The iron hydroxide precipitates were filtered through 0.4 µm Nu-cleopore filters and stored in polyethylene vials for later analysis in the labo-ratory. Particulate chromium was also obtained by filtering unaltered samplesthrough 0.4 µm filters. In the laboratory the iron hydroxide co-precipitateswere dissolved in 6 N distilled hydrochloric acid and analysed by flamelessatomic absorption. The limit of detection of this method is about 0.1 to 0.2 nM.Precision is about 5%.

2.11.2Organic Forms of Chromium

In the determination of the two oxidation states of chromium, the calculationof one oxidation state by differencing presupposes that the two oxidation statesin question were statistically the only contributors to the total concentration.Because of this, contributions from other possible species such as organic com-plexes were generally not considered. However, it has been suggested [39] thatthis presumption may not be warranted and that contributions from organ-ically bound chromium should be considered. This arises from the reportedpresence in non-saline waters of dissolved organic species that form stablesoluble complexes with chromium and may not be readily amenable to de-termination by procedures commonly in use. The results of research into thevalency of chromium present in seawater has not always been consistent. Forinstance, Grimaud and Michard [40] reported that chromium (III) predom-inates in the equatorial region of the Pacific Ocean, whereas Cranston andMurray [35] found that practically all chromium is in the hexavalent state inthe north-east Pacific. Organic chromium (III) complexes may be formed un-der the conditions prevailing in seawater as well as inorganic chromium (III)and chromium (VI) forms. Inconsistencies in earlier research may therefore bedue at least partly to the fact that the possibility of organic chromium specieswas ignored [41, 42].

Nakayama et al. [38] described a method for the determination of chro-mium (III), chromium (VI), and organically bound chromium in seawater.They found that seawater in the Sea of Japan contained about 9 ×10–9 Mdissolved chromium. This was apportioned approximately as 15% inorganicchromium (III), 25% inorganic chromium (VI), and 60% organically boundchromium.

These workers studied the co-precipitation behaviours of chromium specieswith hydrated iron (III) and bismuth oxides.


The collection behaviour of chromium species was examined as follows.Seawater (400 ml) spiked with 10–8 M chromium (III), chromium (VI), andchromium (III) organic complexes labelled with 51Cr was adjusted to thedesired pH by hydrochloric acid or sodium hydroxide.

An appropriate amount of hydrated iron (III) or bismuth oxide was added;the oxide precipitates were prepared separately and washed thoroughly withdistilled water before use [43]. After about 24 h, the samples were filteredon 0.4 µm Nuclepore filters. The separated precipitates were dissolved withhydrochloric acid and the solutions obtained were used for γ -activity mea-surements. In the examination of solvent extraction, chromium was measuredby using 51Cr, while iron and bismuth were measured by electrothermal AAS(EAAS). The decomposition of organic complexes and other procedures werealso examined by EAAS.

Collection of Chromium (III) and Chromium (VI) with Hydrated Iron (III)or Bismuth Oxide

Only chromium (III) co-precipitates quantitatively with hydrated iron (III)oxide at the pH of seawater, around 8. In order to collect chromium (VI)directly without pre-treatment, e.g., reduction to chromium (III), hydratedbismuth oxide, which forms an insoluble compound with chromium (VI) wasused. Chromium (III) is collected with hydrated bismuth oxide (50 mg per400 ml seawater). Chromium (VI) in seawater is collected at about pH 4 andchromium (VI) is collected below pH 10. Thus both chromium (III) and chro-mium (VI) are collected quantitatively at the pH of seawater, i.e., around 8.

Collection of Chromium (III) Organic Complexes with Hydrated Iron (III)or Bismuth Oxide

The percentage collection of chromium (III) with hydrated iron (III) oxide maydecrease considerably in the neutral pH range when organic materials capableof combining with chromium (III), such as citric acid and certain amino acids,are added to the seawater [41]. Moreover, synthesised organic chromium (III)complexes are scarcely collected with hydrated iron (III) oxide over a wide pHrange [41].

As it was not known what kind of organic matter acts as the major ligandfor chromium in seawater, Nakayama et al. [38] used ethylene diaminetetra-acetic acid (EDTA) and 8-quinolinol-4-sulfuric acid to examine the collectionand decomposition of organic chromium species, because these ligands formquite stable water-soluble complexes with chromium (III), although they arenot actually present in seawater. Both of these chromium (III) chelates arestable in seawater at pH 8.1 and are hardly collected with either of the hydratedoxides. The organic chromium species were then decomposed to inorganic

2.11 Chromate and Dichromate 51

chromium (III) and chromium (VI) species by boiling with 1 g ammoniumpersulfate per 400 mL seawater acidified to 0.1 M with hydrochloric acid. Ironand bismuth, which would interfere in AAS, were 99.9% removed by extractionfrom 2 M hydrochloric acid solution of 5% tri-iso-octylamine. Chromium (III)remained almost quantitatively in the aqueous phase in the concentrationrange 10–9 –10–6 M, whether or not iron or bismuth was present. However,as about 95% of chromium (VI) was extracted by the same method, sampleswhich may contain chromium (VI) should be treated with ascorbic acid beforeextraction so as to reduce chromium (VI) to chromium (III).

When the residue obtained by the evaporation of the aqueous phase afterextraction was dissolved in 0.1 M nitric acid and the resulting solution wasanalysed by EAAS, a negative interference seemingly due to residual organicmatter was observed. This interference was successfully removed by digestingthe residue on a hot plate with 1 ml concentrated hydrochloric acid and 3 ml ofconcentrated nitric acid. This process had the advantage that the interferenceof chloride in the AAS was eliminated during the heating with nitric acid.

Results reported by various workers for chromium concentrations in sea-water are listed in Table 2.1. In most of these methods, co-precipitation with

Table 2.1. Data from the literature on the chromium contents of seawater (from author’sown files)

Reference Date Locations Methods Concentration(µg/l)

Ishibashi andShigematsu [44]

1950 Japanese coastal Al (OH)3 co ppt Cr(III) 0.01

Chuecas andRiley [45]

1966 British Coastal Fe(OH)3 co ppt acidicreduction

Cr(III) 0.46Cr(VI) 0.01

Fukai [48, 49] 1967 Mediterranean Fe(OH)3 co ppt acidicreduction

Cr(III) 0.02–0.25Cr(VI) 0.05–0.38

Kuwamoto andMurai [46]

1970 Pacific Ocean BiOH (NO3)2 co pptacidic reduction

Cr(III)? Cr(III)Cr(VI) 0.13–0.90Cr(VI)? OrgCr0.07–0.27

Grimaud andMichard [40]

1974 Pacific Ocean Fe(OH)3 co pptFe(OH)3 reduction

Cr(III) 0.24–0.52Cr(VI) 0.03–0.11

Yamamoto et al.[49]

1974 Pacific Ocean Al(OH)3 co pptacidic reduction

Total 0.06–0.96Cr(III) 17–99%

Cranston andMurray [35]

1978 Pacific Ocean Fe(OH)3 co pptFe(OH)3 reduction

Cr(III) 0.005Cr(VI) 0.15

Nakayama et al.[38]

1981 Pacific Oceanand Japan Sea

Cr(III) 0.06Cr(VI) 0.14OrgCr 0.26Total 0.46


hydrated iron (III) oxide was used to separate chromium (III) from chromi-um (VI) and the chromium (VI) concentration was subsequently determinedby suitable reduction of chromium (VI) to chromium (III) before a further co-precipitation. In others, hydrated iron (II) oxide served as both reductant andcarrier. Isibashi and Shigematsu [44] used co-precipitation with aluminiumhydroxide and did not employ reduction, so that the value reported most likelycorresponds to inorganic chromium (III) alone; in fact, the value for inorganicchromium (III) obtained by Nakayama et al. [38] is in remarkable agreement.

In Chuecas and Riley’s study [45] the samples were stored for a long timeunder acidic conditions before analysis, so that chromium (VI) could havebeen reduced to chromium (III) and any organic chromium dissociated, withthe result that all chromium species would have been determined as chromi-um (III). When a sample is reduced under acidic conditions, organic chromiumis likely to partly dissociate, initially increasing the apparent concentration ofchromium (VI). When the analytical procedure described earlier [40, 46] wasre-examined, the value for chromium (III) was found actually to be the sumof chromium (III) and chromium (VI), while the value for chromium (VI)was partly organic chromium. For the same reason the chromium (VI) valuesdetermined by Fukai [47], Fukai and Vas [48], and Yamamoto et al. [49] prob-ably include organic chromium species. When iron (II) precipitate is used,there seems to be little chance of determining the organic chromium speciesas chromium (VI). The value for chromium (VI) reported by Cranston andMurray [35] agrees quite well with the value for chromium (VI) reported bythem, although the value for chromium (III) is lower. The results obtainedby Grimaud and Michard [40] for chromium (III) differ considerably, but thediscrepancies cannot be discussed because details of the analytical procedurewere not given. It seems reasonable to conclude that the inconsistency of pastresults concerning the dominant chromium species and the total chromiumconcentration in seawater can be attributed, at least partly, to the fact that thepresence of organic chromium species was not considered properly.

Mullins [37] has described a procedure for determining the concentrationsofdissolvedchromiumspecies in seawater.Chromium(III) andchromium(VI)separated by co-precipitation with hydrated iron (III) oxide and total chromi-um are determined separately by conversion to chromium (VI), extractionwith ammonium pyrrolidine diethyl dithiocarbamate into methyl isobutylketone, and determination by AAS. The detection limit is 40 ng/l chromium.The dissolved chromium not amenable to separation and direct extractionis calculated by difference. In waters investigated, total concentrations wererelatively high (1–5 µg/l), with chromium (VI) the predominant species in allareas sampled with one exception, where organically bound chromium wasthe major species.

Ahern et al. [50] have discussed the separation of chromium in seawater.The method involved co-precipitation of trivalent and hexavalent chromium,separately, from samples of surface seawater, and determination of the chromi-

2.12 Fluoride 53

um in the precipitates and particulate matter by thin-film X-ray fluorescencespectrometry. An ultraviolet irradiation procedure was used to release boundmetal. The ratios of labile trivalent chromium to total chromium were in therange 0.4–0.6, and the totals of labile tri- and hexavalent chromium were in therange 0.3–0.5 µg/l. Bound chromium ranged from 0 to 3 µg/l, and represented0–90% of total dissolved chromium. Acidification of the samples in the usualmanner for the determination of trace metals altered the proportion of trivalentto hexavalent chromium.

2.12Fluoride


Fletsch and Richards [51] determined fluoride in seawater spectrophotometri-cally as the cerium alizarin complex. The cerium alizarin complex and chelatewas formed in 20% aqueous acetone at pH 4.35 (sodium acetate buffer) and,after 20–60 min, the extinction measured at 625 nm (2.5 cm cell) against wa-ter. The calibration graph was rectilinear for 8–200 µg/l fluoride; the meanstandard deviation was ±10 µg/l at a concentration of 1100 µg/l fluoride.

Spectrophotometric procedures based on the lanthanum alizarin complexhave also been described [52].

2.12.2Ion Selective Electrodes

Ion selective electrodes are emerging as a method of preference for the deter-mination of fluoride in seawater [53–58].

Anfalt and Jagner [57] measured total fluoride ion concentration by meansof a single-crystal fluoride selective electrode (Orion, model 94-09). Samplesof seawater were adjusted to pH 6.6 with hydrochloric acid and were titratedwith 0.01 M sodium fluoride with use of the semi-automatic titrator describedby Jagner [28]. Equations for the graphical or computer treatment of theresults are given. Calibration of the electrode for single-point potentiometricmeasurements at different seawater salinities is discussed.

Rix et al. [58] have commented that earlier published potentiometric meth-ods for determining fluoride in seawater appear to be unnecessarily compli-cated in several respects. Reagents, such as a total ionic strength adjustmentbuffer (TISAB), are added to the sample at high concentrations to adjust thepH and ionic strength and to liberate the majority of fluoride bound in metalcomplexes. The possibility of introducing solution contamination when usinghigh concentrations of TISAB is always a potential hazard and, if this stepcan be avoided, this will obviously be advantageous. Furthermore, a signif-icant drawback with most earlier methods was the difficulty of establishing


an acceptable or representative matrix in which to prepare standards, as theactual matrix is often unknown in non-saline waters. This aspect of fluoridedetermination in seawater has received considerable attention, and extensiveand elaborate procedures have been used to match the sample and standardmatrices [53].

The method of standard additions is frequently employed in environmentalanalysis for samples of variable matrix [59]. The method of data treatmentused by Rix et al. [58] is similar to that originally described by Gran [2] for theexact determination of potentiometric end-points. Brand and Rechnitz [60],Baumann [61], and Craggs et al. [62] have described procedures employing themethod of standard additions to selective ion electrode potentiometry, but nodata are available for non-saline waters using this direct method except for thework of Liberti and Mascini [63], who determined fluoride in mineral waters.The procedure used by Rix et al. [58] is exceedingly simple, and the method ofdata treatment provides a linear plot, the slope and intercept of which give inde-pendent measures of the original concentration of fluoride in the sample. Totalfluoride concentration is determined despite the fact that the activity of thefluoride ion is the parameter monitored by the fluoride ion selective electrode.

The concentration of fluoride in seawater is approximately 1.4 ppm or7 ×10–5 mol/l and the fluoride electrode has been shown to give a Nerns-tian response at this level – and indeed three orders of magnitude lower thanthis level [64].

It has been well established that the fluoride electrode is highly selectivein its response to the activity of fluoride, with the only common interferantbeing hydroxide [65]. Nevertheless, although seawater has a uniform pH of 8,possible hydroxide interference was shown not to be a problem.

Complexation of fluoride by metal ions in seawater has previously beenovercome by the addition of TISAB solution. The reagent is presumed torelease the bound fluoride by preferential complexation of the metal ionswith EDTA type ligands present in the TISAB. Examination of the metal ionspresent in seawater [66, 67] suggests that magnesium is the major speciesforming fluoride complexes. Theoretical calculations demonstrate that eventhis species is unlikely to interfere.

Results for seawaterof 35.21%salinity arepresented inTable 2.2.Theaveragevalue of 1.35 ± 0.05 mg/l is in excellent agreement with previously publisheddata [53, 54, 68, 69].

The first result in Table 2.2 is the value found directly. The next set of threeresults shows that recovery of fluoride after deliberate addition to the spikedsamples is excellent. The fifth result is that obtained from an acidified sample.The sixth, seventh, eighth, andninth sets ofdata show that the expectedfluorideconcentration is still obtained after deliberate addition of aluminium or iron inthe form of their alums. Aluminium (III) and iron (III) form very strong fluo-ride complexes [70] and, provided that sufficient time is allowed for equilibra-tion (as noted by Baumann [64] for very low fluoride concentrations), total flu-

2.12 Fluoride 55

Table2.2. Determinationoffluoride in seawaterunderavarietyof conditions todemonstratethe reliability of the method [58]

Sample [F] found (mg/l) Conditions

Point Lonsdale,Victoria, Australia

1 1.40 ± 0.051.35 ± 0.05∗

pH = 8

2 1.30 ± 0.05 1.56 ppm spike added. [F]T determined =2.80 ± 0.05 ppm

3 1.40 ± 0.05 4.0 ppm spike added [F]T determined = 5.40±0.05 ppm

4 1.30 ± 0.05 8.6 ppm spike added [F]T determined = 9.90±0.05 ppm

5 1.28 ± 0.05 pH 5.50 (HCl added)6 1.36 ± 0.05 5 mg potassium alum added/100 cm3 seawater

≈ 2.7 ppm Al+37 1.43 ± 0.10 50 mg potassium alum added/100 cm3 seawater

≈ 27 ppm Al+3 . Equilibration time of electrodevery long.

8 1.39 ± 0.051.37 ± 0.05∗

50 mg potassium alum and 3 g ammonium cit-rate added/100 cm3 seawater.

9 1.30 ± 0.05 37 mg ferric alum added to 100 cm3 seawaterand pH adjusted to 3 by adding a few drops ofglacial acetic acid solution ≈ 40 ppn Fe3+.

10 1.37 ± 0.05 50 cm3 TISAB added/100 cm3 seawater11 1.28 ± 0.10 Calculated via addition of TISAB and calibra-

tion versus synthetic seawater.Synthetic 0.0068 ± 0.0008

0.0053 ± 0.0008pH = 5.8pH = 4.3

∗ This determination was performed approximately three weeks after the first.

oride could still be determined accurately by the method of standard additions,even though theartificial concentration levelsof aluminiumand ironwere threeorders of magnitude greater than those normally occurring in seawater [66].This observation is noteworthy, since the work of Liberti and Mascini [63]suggests that the direct method cannot be used for fluoride in the presence ofhigh relative concentrations of aluminium (III) or iron (III) aquo species.

This complication is probably avoided in seawater (even with relatively highconcentrations of aluminium (III) and iron (III)) by the large excess of chlorideions, which partially complex with any aluminium (III) or iron (III) present toproduce chlorocomplexes of the metals, thus sequestering their influence onthe fluoride concentration. Data obtained after adding TISAB to release boundfluoride and using the calibration method further confirm the validity of thedirect method. The data in Table 2.2 demonstrate the absence of hydroxideion interference caused by metal ion complexation, and indicate that totalrather than free fluoride is determined by the method of standard additions.The slope of the linear portion of the graph was 58.8 mV per decade change


in fluoride ion concentration, and this value was used in the calculations offluoride concentration in seawater. At lower levels of fluoride (< 10–6 mol/l),it might be assumed that the electrode behaves in a non-Nernstian fashion.However, calculations of the kind suggested by Rix et al. [58] reveal that thedeviation from linearity can satisfactorily be attributed to residual fluorideresulting either from reagent contamination or from the electrode itself, ata concentration of about 3 ×10–7 mol/l.

The level of fluoride in seawater is at least two orders of magnitude higherthan this residual level, hence contamination does not significantly influencethe results.

Ke and Regier [71] have described a direct potentiometric determinationof fluoride in seawater after extraction with 8-hydroxyquinoline. This proce-dure was applied to samples of seawater, fluoridated tap-water, well-water, andeffluent from a phosphate reduction plant. Interfering metals, e.g., calcium,magnesium, iron, and aluminium were removed by extraction into a solu-tion of 8-hydroxyquinoline in 2-butoxyethanol-chloroform after addition ofglycine–sodium hydroxide buffer solution (pH 10.5 to 10.8). A buffer solu-tion (sodium nitrate–1,2-diamino-cyclohexane–N,N,N ′,N ′–tetra-acetic acid–acetic acid; pH 5.5) was then added to adjust the total ionic strength and thefluoride ions were determined by means of a solid membrane fluoride-selectiveelectrode (Orion, model 94-09). Results were in close agreement with and morereproducible than those obtained after distillation [72]. Omission of the ex-traction led to lower results. Four determinations can be made in one hour.

2.12.3Photoactivation Analysis

Photoactivation analysis has also been used to determine fluoride in seawa-ter [73]. In this method a sample and simulated seawater standards containingknown amounts of fluoride are freeze-dried, and then irradiated simultane-ously and identically, for 20 min, with high-energy photons. The half-life of18F (110 min) allows sufficient time for radiochemical separation from the sea-water matrix before counting. The specific activities of sample and standardsbeing the same, the amount of fluoride in the unknown may be calculated.The limit of detection is 7 ng fluoride, and the precision is sufficient to permitdetection of variations in the fluoride content of oceans. The method can beadapted for the simultaneous determination of fluorine, bromine, and iodine.


Fluoride has been determined in seawater in amounts down to 8 ng/ml bya method based on the formation of AlF in an electrothermal graphite furnace,followed by molecular absorption at 227.45 nm [74].

2.13 Formate 57

2.13Formate

There is considerable interest in the role of formic acid and other volatilefatty acids in the early diagnosis of organic matter in lacustrine and marinesediments. Formic acid is an important fermentation product or substrate formany aerobic and anaerobic bacteria and for some yeasts. In the atmosphere,formic acid is an important product in the photochemical oxidation of organicmatter.

Despite its potential importance, formic acid has proven difficult to quan-tify at submicromolar levels in non-saline water samples. Formidable ana-lytical difficulties are associated with its detection in highly saline samples.Ion exclusion, anion exchange, and reversed-phase high performance liquidchromatography techniques based on the direct detection of formic acid inaqueous samples are prone to interferences (especially from inorganic salts)that ultimately limit the sensitivity of these methods.

A potentially more sensitive and selective approach involves reaction offormic acid with a reagent to form a chromophore or fluorophore, followed bychromatographic analysis. A wide variety of alkylating and silylating reagentshave been used for this purpose. Two serious drawbacks to this approach arethat inorganic salts and/or water interfere with the derivatisation reaction, andthese reactions are generally not specific for formic acid or other carboxylicacids. These techniques are prone to errors from adsorption losses, contami-nation, and decomposition of the components of interest. Enzymic techniques,in contrast, are ideal for the analysis of non-saline water samples, since they arecompatible with aqueous media and involve little or no chemical or physicalalterations of the sample (e.g., pH, temperature).

2.13.1High Performance Liquid Chromatography (HPLC)

As a consequence of the previous considerations Kieber et al. [75] have devel-oped an enzymic method to quantify formic acid in non-saline water samplesat sub-micromolar concentrations. The method is based on the oxidationof formate by formate dehydrogenase with corresponding reduction of β-nicotinamide adenine dinucleotide (β-NAD+) to reduced β-NAD+(β-NADH);β-NADH is quantified by reversed-phase high performance liquid chromatog-raphy with fluorimetric detection. An important feature of this method is thatthe enzymic reaction occurs directly in aqueous media, even seawater, anddoes not require sample pre-treatment other than simple filtration. The reac-tion proceeds at room temperature at a slightly alkaline pH (7.5–8.5), and isspecific for formate with a detection limit of 0.5 µm (S/N = 4) for a 200 µl injec-tion. The precision of the method was 4.6% relative standard deviation (n = 6)for a 0.6 µM standard addition of formate to Sargasso seawater. Average re-


coveries of 2 µM additions of formate to seawater were 103%. Intercalibrationwith a Dionex ion chromatographic system showed an excellent agreement of98%. Concentrations of formate present in natural samples ranged from 0.2 to0.8 µM for Biscayne Bay seawater.

2.14Hypochlorite


Williams and Robertson [76] have described a simple inexpensive methodfor determining reactive chlorine in non-saline waters. It involves additionof bromine, which is oxidised by the reactive chlorine in the sample, andwhich in turn brominates fluorescein to give a pink derivative; this can bemeasured visually or spectophotometrically, or the decrease in fluoresceincan be measured fluorimetrically. Potential applications of the method areindicated.

The basis of this method is that when normal seawater is chlorinated at theusual levels of 1 to 10 mg/l of chloride, the bromine in seawater (8.1 ×10–4 M,65 mg/l at salinity = 35‰) is rapidly and quantitatively oxidised to BrO– andHBrO. If 50 mg/l of bromide is added to distilled or fresh waters containingHClO plus ClO–, then HBrO plus BrO– are both formed. The HBrO plus BrO–

will in turn rapidly brominate fluorecein (9-[o-carboxyphenyl]-6-hydroxy-3-isoxanthenone) to give the pink tetrabromo derivative eosin yellow (2,4,5,7-tetrabromo-9-[o-carboxyphenyl]-6-hydroxy-3-isoxanthenone), provided themolar ratio of bromide to fluorescein is 4:1. The resultant increase in eosincan be measured visually or spectrophotometrically, and the decrease in fluo-roscein measured fluorometrically. If the molar ratio of bromide to fluorosceinis < 4:1, then the mono-, di-, and tri-bromo derivatives are formed repro-ducibly. These derivatives have extinction coefficients close to eosin and areaccounted for in the standardisation.

2.15Iodate


Truesdale [77] has described autoanalyser procedures for the determinationof iodate and total iodine in seawater. The total iodine content of seawater(approximately 50–60 µg/l) is believed to be composed of iodate (30–60 µg/lof I) and iodine–iodine (0–20 µg/l) with perhaps a few µg/l of organically

2.15 Iodate 59

bound iodine. In both of the methods, the iodine species of interest are firstconverted to iodate. Then the iodate is reacted with acid and excess iodide togive iodonium ions, I–

3, which are detected spectrophotometrically at 350 nm.In the total iodine procedure, a pre-oxidation step is therefore required; herebromine water was chosen. Interference from nitrite ions, which in acid alsooxidise iodide to iodonium ions, is suppressed by sulfamic acid, which destroysthe nitrite ions.

Originally the objective of this work was only the automatic iodate–iodineprocedure. This was needed to give the extra precision called for in an earlierstudy [78], which suggested that most of the observed variation in iodate–iodine results for the deeper waters (> 200 m) of the oceans is due to analyticalimprecision. However, in addition to the original objective the procedure fortotal-iodine has developed. This development stemmed from the need to testthe likelihood of iodonium ions, produced in the iodate procedure, being re-duced by substances occurring naturally in seawater. This problem was definedwhen Truesdale [79] showed that molecular iodine is reduced rapidly in someseawaters. He found that the oxidising capacity of 260 µg/l of iodine–iodineadded to a filtered coastal seawater disappeared within 30 min. The effect ofsuch a process on the iodate method would be to lower the observed iodateconcentrations. Further, the effect, if it occurred, would be expected to have itsmaximum effect in coastal and oceanic surface waters where primary produc-tivity is highest; these waters also appear to contain the lowest recorded iodateconcentrations. To test for the iodonium-ion reduction, a pre-oxidation stepincluding iodine–iodine was incorporated in the analytical method for iodate–iodine. Having accomplished the iodine water pre-oxidation step successfully,it was logical to attempt the bromine water pre-oxidation and thereby producethe total-iodine method.

Methods for the following three determinations were described by Trues-dale [77]:

(1) Determination of iodate without pre-oxidation

Iodate in the buffered sample is reacted with sulfamic acid (to destroy nitrite)and potassium iodide to produce the iodonium ion I–

3, which is determinedspectrophotometrically at 350 nm.

(2) Determination of iodate with pre-oxidation

Iodine water is added to an acetic acid sodium acetate buffered sample toreoxidise to iodate any iodine-containing substances produced by reduction ofiodate by naturally occurring reducing substances present in the sample. Totaliodate (i.e., iodatepresent in theoriginal sampleas iodateplus additional iodateproduced by iodine water treatment) is then reacted with phenol solution at


pH 5.4 to destroy excess free iodine, and then with sulfamic acid and potassiumiodide toproduce the iodoniumion,which is estimated spectrophotometricallyat 350 nm.

This determination will test for the presence of naturally occurring reducingagents in seawater which by their action on iodonium ions could lead to anunderestimate in iodate concentration. (The use of the method on anoxicwaters containing sulfide is a prime example of when this precaution shouldbe taken.)

(3) Determination of total iodine with pre-oxidation

This procedure is the same as that described in (2) except that the iodine wateris replaced by bromine water, and the buffer contains added sodium bromide,[i.e, sodium bromide–acetic acid–sodium acetate instead of the acetic acid–sodium acetate buffer used in method (2)].

In all three methods a blank is obtained. To ascertain the blank, excess sodiumthiosulfate is added to the potassium iodide reagent at a concentration of4.0 ×10–14 mol/l. Samples were re-analysed and the appropriate blank sub-tracted from the sample signal.

Variations in the salinity of the sample have very little effect on the accuracyof the results obtained in all three methods provided that the difference insalinity between the samples and the standards does not exceed 2.5‰.

Determinations of iodate without pre-oxidation in Pacific seawater by theprevious method gave a mean result of 583 µg/l with a standard deviationof 0.23 µg/l. For samples containing between 40 and 60 µg/l, standard devia-tions of 0.19 µg/l (iodate method with pre-oxidation), 0.12 µg/l (iodate methodwithout pre-oxidation), and 0.43 µg/l (total iodine method) were obtained.

A set of Pacific open-ocean samples were analysed for iodate–iodine usingboth the procedure which incorporates pre-oxidation with iodine water andthat which does not. Also, in a similar exercise total iodine was determinedusing both the method that incorporates pre-oxidation with bromine waterand the catalytic method using the reaction between Ce(IV) and As(III) [81].Variance tests showed that differences between either replicates or methodswas not significant.

Truesdale and Smith [80] also carried out a comparative study of the de-termination of iodate in open ocean, inshore Irish seawaters and waters fromthe Menai Straits, using the spectrophotometric method (with and withoutpre-oxidation using iodine water) and also by a polarographic method [82].

Figure 2.3 shows the results obtained in a comparision of these methodson a range of deep-sea and offshore samples. The line of gradient 1.0 on eachdiagram shows the result which would have occurred had agreement beenobtained. The Student t-test showed that in both exercises the colorimetricmethod with iodine water treatment yielded higher values than that without

2.15 Iodate 61

Figure 2.3. The distribution of sample concentrations encountered in the comparisons ofthe colorimetric method without iodine with (a) the polarographic method and (b) thecolorimetric method with iodine water. For convenience the line of the unit gradient isshown in both cases. In each case the precision (95% confidence level) for each measurementis shown for the most concentrated sample. In (b) the broken line divides on shore (�) andoffshore (�) samples. From [80]

iodine water treatment, and the polarographic method yielded, on average,a concentration lower than thatobtainedby thecolorimetricprocedurewithoutiodine water.

Schnepfe [83] has described yet another procedure for the determinationof iodate and total iodine in seawater. To determine total iodine 1 ml of 1%aqueous sulfamic acid is added to 10 ml seawater which, if necessary, is filteredand then adjusted to a pH of less than 2.0. After 15 min, 1 ml sodium hydroxide(0.1 M) and 0.5 ml potassium permanganate (0.1 M) are added and the mixtureheated on a steam bath for one hour. The cooled solution is filtered and theresidue washed. The filtrate and washings are diluted to 16 ml and 1 ml ofa phosphate solution (0.25 M) added (containing 0.3 µg iodine as iodate perml) at 0 ◦C. Then 0.7 ml ferrous chloride (0.1 M) in 0.2% v/v sulfuric acid, 5 mlaqueous sulfuric acid (10%) – phosphoric acid (1:1) are added at 0 ◦C followedby 2 ml starch–cadmium iodide reagent. The solution is diluted to 25 ml andafter 10–15 min the extinction of the starch–iodine complex is measured ina –5 cm cell. To determine iodate the same procedure is followed as is describedpreviously except that the oxidation stage with sodium hydroxide – potassiumpermanganate is omitted and only 0.2 ml ferrous chloride solution is added.A potassium iodate standard was used in both methods.

The total iodine procedure is claimed to be relatively free from interferenceby foreign ions. The iodate procedure is subject to interference by bromateand sulfite ions. This method is claimed to be capable of determining down to0.1 µg iodine in the presence of 500 mg chloride ion and 5 mg of bromide ion.


2.16Iodide

Iodine species in seawater exist as iodide (I–) and iodate (IO–3). Iodide, which

is thermodynamically unstable in oxygenated water, is usually a minor speciesin seawater compared to iodate, and total inorganic iodine concentration isapproximately in the range 50–60 µg/l [84–95]. However, iodide concentrationranges from 30 µg/l near the shore and in ocean surface and bottom watersto below 1 µg/l (deep ocean water). Iodine is an essential micronutrient formany organisms. Iodide in seawater is produced by biologically mediated re-duction of iodate [96] and is also produced under reduction conditions. Thus,the distribution of iodide and iodate gives clues for understanding the marineenvironment [87, 88, 93, 95]. Iodide has been determined mainly by differencebetween total inorganic iodine (I– andIO–

3)andIO–3 iodatehasbeendetermined

by differential pulse polarography [88,89,91,94–97] and the spectrophotomet-ric method after conversion of iodate to I–

3 in acidic solution [83, 92]. Totalinorganic iodine was determined after converting iodide to iodate by usingchemical or photochemical means [82, 84, 85, 87, 91, 93, 94]. However, as thedetection limit of iodide by the difference technique is not very good (for ex-ample, 1.3–2.5 µg/l for differential pulse polarography [88]), it is desirable todetect iodide directly.

To date, a few methods have been proposed for direct determination oftrace iodide in seawater. The first involved the use of neutron activation anal-ysis (NAA) [86], where iodide in seawater was concentrated by strongly basicanion-exchange column, eluted by sodium nitrate, and precipitated as pal-ladium iodide. The second involved the use of automated electrochemicalprocedures [90]; iodide was electrochemically oxidised to iodine and wasconcentrated on a carbon wool electrode. After removal of interference ions,the iodine was eluted with ascorbic acid and was determined by a polishedAg3SI electrode. The third method involved the use of cathodic strippingsquare wave voltammetry [92] (See Sect. 2.16.3). Iodine reacts with mercuryin a one-electron process, and the sensitivity is increased remarkably by theaddition of Triton X. The three methods have detection limits of 0.7 (250 mlseawater), 0.1 (50 ml), and 0.02 µg/l (10 ml), respectively, and could be ap-plied to almost all the samples. However, NAA is not generally employed. Thesecond electrochemical method uses an automated system but is a specialapparatus just for determination of iodide. The first and third methods aretime-consuming.


Iodide in seawater can be determined by the procedure described below, whichis capable of determining down to 0.1 mg/l iodide [97]. In this method the

2.16 Iodide 63

sample is strongly acidified with hydrochloric acid titrated with 0.00125 Mpotassium iodate solution, which converts iodide via iodine monochloride:

KIO3 + 2KI + 6HCI = 3ICl + 3H2O

The end-point, which occurs with the complete conversion of iodide to iodinemonochloride, is indicated by the disappearance of the violet iodine colourfrom a chloroform layer present in the titration flask.


Sugawara [98] has described a spectrophotometric method for the determi-nation of iodide in seawater. Various workers [99, 100] have modified thisprocedure.

Matthews and Riley [99] preconcentrated iodide by co-precipitation withchloride ions. This is achieved by adding 0.23 g silver nitrate per 500 ml ofseawater sample. Treatment of the precipitate with aqueous bromine and ul-trasonic agitation promote recovery of iodide as iodate which is caused to reactwith excess iodide under acid conditions, yielding I–

3. This is determined eitherspectrophotometrically or by photometric titration with sodium thiosulfate.Photometric titration gave a recovery of 99.0 ± 0.4% and a coefficient of vari-ation of ±0.4% compared with 98.5 ± 0.6% and ±0.8%, respectively, for thespectrophotometric procedure.

Shizuo [100] allowed the silver halide precipitate obtained in the co-preci-pitation process to stand in contact with the solution for more than 20 h toensure quantitative collection of iodide on the precipitate. He then evaporatedthe oxidised iodate solution to 5–10 ml and again allowed the solution tostand for more than 12 h before the colorimetric determination. There was nointerference from bromine compounds. The errors were then within ±3%.

2.16.3Cathodic Stripping Voltammetry

Luther et al. [92] have described a procedure for the direct determination ofiodide in seawater. By use of a cathodic stripping square-wave voltammetry, itis possible to determine low and sub-nanomolar levels of iodide in seawater,freshwater, and brackish water. Precision is typically ±5% (1σ). The minimumdetection limit is 0.1–0.2 nM (12 parts per trillion) at 180 sec deposition time.Data obtained on Atlantic Ocean samples show similar trends to previouslyreported iodine speciation data. This method is more sensitive than previ-ous methods by 1–2 orders of magnitude. Triton X-100 added to the sampleenhances the mercury electrode’s sensitivity to iodine.


2.16.4Ion Chromatograpy

This is a simple method for determining trace iodide in seawater. Two factorsare essential to attain good sensitivity:

1. Separation of iodide from an excess of anions in seawater.2. Highly sensitive detection of iodide.

Iodide was determined by an iodide-selective electrode (Ag2S/AgI) after otheranions were separated by a rhodium nitrate element [101]. However, the elec-trode that was stabilised by 0.5 µm iodide responded to chloride ions in sea-water, and the detection limit of iodide was 22 µg/l.

Ito and Sumahara [102] showed that 0.1 M sodium chloride eluent was ef-fective for the separation of iodide in seawater, due to faster elution of anexcess of anions in seawater and slow elution of iodide with hydrophobic-ity on conventional low capacity anion exchange columns. In a more recentwork Ito [103] has described a simple and highly sensitive ion chromato-graphic method with ultraviolet detection for determining iodide in seawater.A high-capacity anion-exchange resin with polystyrene–divinylbenzene ma-trix was used for both preconcentration and separation of iodide. Iodide inartificial seawater (salinity: 35‰) was trapped quantitatively (98.8 ± 0.6%)without peak broadening on a preconcentrator column, and was separatedwith 0.35 M sodium perchlorate and 0.01 M phosphate buffer (pH 6.1). Onthe other hand, the major anions in seawater, chloride and sulfate ions, werepartially trapped (5–20%) and did not interfere in the determination of io-dide. The detection limit for iodide was 0.2 µg/l for 6 ml of artificial seawater.This method when applied to determination of iodide had a detection limitof 18.3 µg/l.

2.16.5Miscellaneous

Buchberger et al. [104] carried out a selective determination of iodide in brine.The performance of a potentiometric method using an ion-selective electrodeand of liquid chromatography coupled with ultraviolet detection at 230 nmwere compared as methods for the determination of iodide in the presence ofother iodide species. Satisfactory results were obtained from the potentiomet-ric method provided the solution was first diluted tenfold with 5 M sodiumnitrate, and external standards were used. Better reproducibility was, however,achieved with HPLC, provided precautions were taken to prevent reductionof iodine to iodide in the mobile phase, for which extraction of iodine withcarbon tetrachloride prior to analysis was recommended. This was the pre-

2.17 Molybdate 65

ferred method for automated analyses of iodine brines (iodide concentrationsof 30–40 mg/l).

Krishnamoorthy and Iyer [105] have reported a method for determin-ing nanogram levels of iodide in saline water samples containing a largeexcess of interfering chloride ions. The anions are first bound to a strongbase anion exchanger, from which the chloride ion is readily eluted. The io-dide is then eluted with 2 M ammonium nitrate and the iodide is determinedbased on its catalytic effect on the reduction of cerium (IV) by arsenic (III).The method is claimed to have an accuracy comparable to that obtained byNAA.

Varma [106] determined iodide in seawater in amounts down to 2 mg/l bya method based on pre-column derivatisation of the iodide into 4-iodo-2,6-di-methyl phenol. An ultraviolet detector was employed.

2.17Molybdate


A limited amount of work has been carried out on the determination of molyb-denum in seawater by AAS [107–109] and graphite furnace atomic absorptionspectrometry [110]. In a recommended procedure a 50 ml sample at pH 2.5 ispreconcentrated on a column of 0.5 g p-aminobenzylcellulose, then the columnis left in contact with 1 mol/l ammonium carbonate for 3 h, after which three5 ml fractions are collected. Finally, molybdenum is determined by AAS at312.2 nm with use of the hot-graphite-rod technique. At the 10 mg/l level thestandard deviation was 0.13 µg.

2.18Nitrate


Spencer and Brewer [111] have reviewed methods for the determination ofnitrate in seawater. Classical methods for determining low concentrations ofnitrate in seawater use reduction to nitrite with cadmium/copper [112,116,117]or zinc powder [113] followed by conversion to an azo dye using N-1-naphthyl-ethylenediamine dihydrochloride and spectrophotometric evaluation. Malho-tra and Zanoni [114] and Lambert and Du Bois [115] have discussed theinterference by chloride in reduction-azo dye methods for the determinationof nitrate.


2.18.2Ultraviolet Spectroscopy

Ultraviolet spectrometry has also been used to determine nitrates. In a methodfor determining high levels of nitrate described by Mertens and Massart [118]the sample, diluted to contain 0.5–1 mg/l nitrate is acidified, filtered througha 0.5 µm filter, and the extinction measured against a blank at 210–220 nm.The concentration of nitrate is obtained from a calibration graph. Interferencefrom chloride, bromide, organic matter, carbonate, bicarbonate, and nitrite islargely removed by using as blank a solution prepared by boiling 10 ml of thesample with 0.5 g Raney nickel for 30 min and then stirring for 90 min at 90 ◦C,to reduce nitrate to ammonia. The extinction coefficient is 8500 at 210 nm and4100 at 220 nm. Beer’s law is valid up to 10 mg/l at 210 nm and up to 15 mg/l at220 nm. The standard deviations for a sample containing 17.4 mg/l of nitratewere 0.3 and 0.5 mg/l at 210 and 220 nm, respectively. No systematic errorswere detected.

Previous ultraviolet methods for determining nitrate have attempted toallow for humic acid interference [119–122]. However, with the exceptionof Morries [122] these methods of allowance are inaccurate at humic acidconcentrations above about 3.5 mg/l. Unfortunately, none of these methodshave attempted to make any allowance for ultraviolet-absorbing pollutantorganic compounds or interfering inorganic ions. Thus their application towater analyses other than for relatively unpolluted fresh waters is open toquestion.

Brown and Bellinger [123] have proposed an ultraviolet technique that isapplicable to both polluted and unpolluted fresh and some estuarine waters.Humic acid and other organics are removed on an ion exchange resin. Bromideinterference in seawater samples can be minimised by suitable dilution of thesample but this raises the lower limit of detection such that only on relativelyrich (0.5 mg/l NO3N) estuarine and inshore waters could the method be used.Chloride at concentrations in excess of 10 000 mg/l do not interfere.

The method is either not affected by or can allow for interference from phos-phate, sulfate, carbonate, bicarbonate, nitrate, coloured metal complexes, am-monia dyes, detergents, phenols, and other ultraviolet-absorbing substances.The method incorporates three features designed to reduce interferences:

1. Humicacid interference is reducedbycarryingoutmeasurementsat 225 nm,a longer wavelength than that used by previous workers (210–220 nm).

2. Removal of inorganic interferences, particularly the removal of bromideinterference in seawater by fivefold dilution of the sample, the removal ofnitrate by addition of sulfamic acid, and the removal of metals by passagethrough Amberlite IR 120 cation exchange resin.

3. Removal of ultraviolet absorbing organics by passage through a specific ionexchange resin such as Amberlite XAD-2.

2.18 Nitrate 67

Table 2.3 comparesnitratedeterminations in thepresenceof various interferingsubstances for this method and for two alternate methods – phenoldisulfonicacid and ion selective electrode methods. In general, the method proposed byBrown and Bellinger [123] is less subject to interference.

The speed of the nitrate ion selective electrode makes its use potentiallyideal for nitrate determinations on a large number of samples. However, theresults fromaddingvarious interferingsubstances (Table2.3) seemtocast somedoubt upon the values obtained in the presence of chloride and bicarbonate,for although the results are precise, they are not accurate, being approximately20–30% high.

Table2.3. The effects of various interferences on the determination of nitrates by three meth-ods. For all determinations the concentration of NO3N was 1.125 mg/l. All concentrationsare given in mg/l [123]

Interference MethodPhenoldisulfonic acid UV method Selective ion electrode

Uncorrected Corrected Uncorrected Corrected Uncorrected Corrected

Chloride:100 0.62 1.07∗ 1.08 1.08 1.55§

1000 1.05

Bicarbonate–Carbonate:100 1.13 1.06 0.87 1.24¥

200 1.231000 1.13

Bicarbonate–Chloride:100 Cl:200HCO3

1.58¶

100 Cl:100HCO3

1.46¶

50 Cl:100HCO3

1.46¶

Nitrate:0.304 1.11 1.48 1.10†

1.500 1.31¶

3.040 1.13 1.06†

Ammonium:1.4 0.9814.0 0.66 1.08 1.13¶

1400.0 1.08

∗ Addition of exact amount of silver sulfate to precipitate chloride with no excess silver.† Addition of sulfamic acid (final concentration in sample approximately 20 mg/l).§ Addition of saturated silver sulfate (1–10 ml sample). Not recalibrated. ¥ Addition of2 M acetic acid to sample (0.09 ml per 30 ml sample). Not recalibrated. ¶ Calibrate forpresence of 3 ml saturated silver sulphate and 0.09 ml of 2 M acetic acid in 30 ml sample.


2.18.3Chemiluminescence Method

Chemiluminescent techniques have been used to determine nanomolar quan-tities of nitrate and nitrite in seawater [124, 125]. This method depends onthe selective reduction of these species to nitric oxide, which is then de-termined by its chemiluminescent reaction with ozone, using a commercialnitrogen oxides analyser. The necessary equipment is compact and suffi-ciently sturdy to allow shipboard use. A precision of ±2 nmol/l is claimed,and an analytical range of 2 nmol/l with analysis rates of 10–12 sampleshourly.

In thismethod[124]nitrate,nitrateplusnitriteornitrite aloneare selectivelyreduced to nitric oxide, which is swept from the sample in a helium carrier gasflow. Nitric oxide is allowed to react with ozone in a nitrogen oxide analyser,where it forms nitrogen dioxide. The return of the nitrogen dioxide to theground state is accompanied by release of a photon, which is detected bya photomultiplier. The integrated output of the photomultiplier over the timethat the nitric oxide is purged from the sample is proportional to the nitritecontent of the sample.

Linear calibration plots are obtained by this procedure covering the rangeup to 1 nmol/l for nitrate and for nitrite.

2.18.4Flow Injection Analysis

Flow injection analysis is another technique that has been applied to the deter-mination of nitrate and nitrite in seawater. Anderson [126] used flow injectionanalysis to automate the determination of nitrate and nitrite in seawater. Thedetection limit of his method was 0.1 µmol/l. However, the sampling rate wasonly 30 per hour which is low for flow injection analysis. Reactions seldomgo to completion in a determination by flow injection analysis [127, 128] be-cause of the short residence time of the sample in the reaction manifold.Anderson selected a relatively long residence time so that the extent of for-mation of the azo dye was adequate to give a detection limit of 0.1 µmol/l.This reduced the sampling rate because only one sample is present at a timein the post-injector column in flow injection analysis. Any increase in reac-tion time causes a corresponding increase in the time needed to analyse onesample.

Johnson and Petty [129] reduced nitrate to nitrite with copperised cadmium,which was then determined as an azo dye. The method is automated by meansof flow injection analysis technique. More than 75 determinations can be madeper hour. The detection limit is 0.1 µmol/l, and precision is better than 1% atconcentrations greater than 10 µmol/l.

2.18 Nitrate 69

2.18.5Continuous Flow Analysis

Hyde and Hill [130] used the copper–cadmium reduction method to deter-mine nitrate in seawater. The construction of a copper–cadmium (50–50 w/w)reactor column for use during continuous flow analysis at sea is described.A 100% yield could be obtained using a 20 cm×3 cm column fitted with grainsof copper–cadmium alloy between 500 and 350 µm in size. The column main-tained its reactivity during three months storage prior to an oceanographiccruise, and during the four week cruise period. Its performance was similar tothat of the Stainton-type cadmium wire column, but it had the advantages ofeasier preparation and easier control of reductor volume.


VandenBerg [131]used this technique todeterminenanomolar levelsofnitratein seawater. Samples of seawater from the Menai Straits were filtered and nitritepresent reacted with sulfanilamide and naphthyl-amine at pH 2.5. The pH wasthen adjusted to 8.4 with borate buffer, the solution de-aerated, and thensubjected to absorptive cathodic stripping voltammetry. The concentration ofdye was linearly related to the height of the reduction peak in the range 0.3–200 nM nitrate. The optimal concentrations of sulfanilamide and naphthyl-amine were 2 mM and 0.1 mM, respectively, at pH 2.5. The standard deviationof a determination of 4 nM nitrite was 2%. The detection was 0.3 nM for anadsorption time of 60 sec. The sensitivity of the method in seawater was thesame as in fresh water.


Tyree and Bynum [132] described an ion chromatographic method for thedetermination of nitrate and phosphate in seawater. The pre-treatment com-prisedvigorousmixingof the samplewith a silver-basedcation-exchange resin,followed by filtration to remove the precipitated silver salt.

Dahiloef et al. [133] have described an ion chromatographic method for thedetermination of nitrate and phosphate in seawater. A small sample volume(20 µl) is needed, and detection limits of 0.5 and 1.0 µM were obtained fornitrate and phosphate, respectively.

2.18.8Bacteriological Method

Sigman et al. [134] have pointed out that nitrate is the predominant formof bioavailable (or “fixed”) nitrogen in the ocean, and the natural isotopic


variations of this species provide an important tool for studying the nitro-gen cycle. Depending on the environment, the 15N/14N ratio of seawater ni-trate can provide information on virtually all of the major transformationsof nitrogen that occur in the ocean, including dinitrogen fixation, uptakeof fixed nitrogen by phytoplankton, nitrification, and denitrification. The18O/16O ratio of nitrate has been studied in freshwater and terrestial sys-tems, and has been shown to provide an additional important constrainton natural processes. Generally, both nitrogen and oxygen isotopic compo-sitions of nitrate have many potential applications in oceanography, hydrol-ogy, and atmospheric chemistry, but natural-abundance isotopic studies ofnitrate have been restricted by analytical limitations, especially in marinesystems.

Methods to measure the nitrogen isotopic composition of nitrate in naturalwaters typically involve the reduction of nitrate to ammonia using diffusionor distillation, reaction to nitrogen gas, and isotopic analysis of the nitro-gen [135]. Methods also exist for the coupled nitrogen and oxygen isotopicanalysis of nitrate in freshwater that are based on the purification of the ni-trate salt, and the direct, high-temperature conversion of nitrate to N2 andCO2 [136–138]. However, weaknesses in the available methods have madesome nitrate isotope investigations difficult and have precluded others. Thepublished N2-based methods for nitrogen isotopic analysis normally requiremicromoles of nitrate-N, which is prohibitive when only millilitres of wa-ter are available. The methods available for the nitrogen isotopic analysis ofseawater nitrate require nitrate concentrations of 2–3 µM or higher [135],largely because of the limited efficiency with which ammonia is extracted bydistillation and diffusion. In addition, these methods have both a significantreagent blank and a blank associated with dissolved organic nitrogen thatvaries with sample type, which can be large [139]. Finally, these methods aretypically labour- or time-intensive. With respect to oxygen isotopic analysis,the weaknesses in the available methods are even more restrictive. In particu-lar, there is no published method for the oxygen isotopic analysis of nitrate inseawater.

Sigman et al. [134] have described a bacterial method for measuring theisotopic composition of seawater nitrate at the natural-abundance level. Themethod is based on the analysis of nitrous oxide gas (N2O) produced quan-titatively from nitrate by denitrifying bacteria. The classical denitrificationpathway consists of the stepwise reduction of nitrate (NO–

3) to nitrite (NO–2),

nitric oxide (NO), nitrous oxide (N2O), and dinitrogen (N2):

NO–3 → NO–

2 → NO → N2O → N2

Each of these steps is carried out by a dedicated enzyme encoded by a distinctgene. There is a rich literature on natural and genetically modified bacterialstrains that lack discrete components of the denitrification pathway [140]. The

2.19 Nitrite 71

method described takes advantage of naturally occurring denitrifiers that lackan active N2O reductase, the enzyme that reduces N2O to N2 [140].

The precision of the method is better than 0.2‰ (1 SD) at concentrationsof nitrate down to 1 µM, and the nitrogen isotopic differences among variousstandards and samples are accurately reproduced. For samples with 1 µMnitrate or more, the blank of the method is less than 10% of the signal size, andvarious approaches may reduce it further.

2.18.9Miscellaneous

Cerda et al. [142] have described a sequential injection sandwich technique forthe simultaneous determination of nitrate and nitrite in seawater.

2.19Nitrite


Spencer and Brewer [144] have reviewed methods for the determination ofnitrite in seawater. Workers at WRc, UK [145] have described an automatedprocedure for the determination of oxidised nitrogen and nitrite in estuar-ine waters. The procedure determines nitrite by reaction with N-1 naphthyl-ethylene diamine hydrochloride under acidic conditions to form an azo dyewhich is measured spectrophotometrically. The reliability and precision ofthe procedure were tested and found to be satisfactory for routine analyses,provided that standards are prepared using water of an appropriate salin-ity. Samples taken at the mouth of an estuary require standards preparedin synthetic seawater, while samples taken at the tidal limit of the estuaryrequire standards prepared using deionised water. At sampling points be-tween these two extremes there will be an error of up to 10% unless thesalinity of the standards is adjusted accordingly. In a modification of themethod, nitrate is reduced to nitrite in a micro cadmium/copper reductioncolumn and total nitrite estimated. The nitrate content is then obtained bydifference.

Matsunaga et al. [146] have also described a similar procedure for thedetermination of nitrite in seawater. To 500 ml of sample is added 10 ml ofsulfanilamide (1% solution) in 2 M hydrochloric acid and 5 ml aqueous N-1-naphthylethylene-diamine hydrochloride (0.1%). After 10 min, 5 ml of aqueousdodecylbenzenesulfonate (0.5%) is added and the mixture extracted for 2 minwith 50 ml of carbon tetrachloride. Then 75 ml of acetone and 10 ml of 0.1 Mhydrochloric acid are added to the separated organic layer and the azo dye


back-extracted into the acid. The extinction of the acid layer is measured at543 nm in a 2 cm cell. The sensitivity for 0.001 extinction in 1 cm cells is 0.7 ngatom nitrite-N per litre.

Gianguzza and Orecchio [147] have carried out comparative trials of var-ious methods for estimating nitrites in seawaters. These workers compareda method using sulfanilic acid/α-naphthylamine complexes with a method us-ing sulfanilamide/N(1-naphthyl) ethylenediamine complexes for the determi-nationofnitrites in salinewaters. The secondmethodhas thegreater sensitivityand lower detection limits. The former method is subject to interference fromchlorides, and this interference can be completely eliminated by the couplingdiazotisation procedure of the latter method.

Anion exchange resins have been used to determine extremely low concen-trations of nitrite down to nanomoles in seawater. Wada and Hattori [148]formed an azo dye from the nitrite using N-1 naphthylethylene diaminedihydrochloride and then adsorbed the dye in an anion exchange resin. Thedye is then eluted from the column with 60% acetic acid and measured spec-trophotometrically at 550 nm.


Fogg et al. [149] used flow injection voltammetry to reduce nitrite at a glassycarbon electrode in acidic bromide or chloride medium and applied themethod to seawater.

2.19.3Isotope Dilution Gas Chromatography

Mass spectrometry has been applied to the determination of nitrite and nitratein seawater. This method produced results that were free from interference byother forms of nitrogen [150].


Absorptive cathodic stripping voltammetry has been used [151, 152] to de-termine nanomolar levels of nitrite in seawater. The nitrite is derivatised bydiazotisation with sulfanilamide and coupled with 1-naphthylamine to forman azo dye. The dye adsorbs onto a mercury drop electrode and its reductionis fully reversible. The concentration of dye is linearly related to concentrationof nitrite in the range 0.3–200 nM. Down to 0.3 nM nitrite can be determinedin seawater for an adsorption time of 60 seconds.

2.20 Nitrate and Nitrite 73

2.20Nitrate and Nitrite


Bajic and Jaselskis [153] described a spectrophotometric method for the deter-mination of nitrate and nitrite in seawater. It included the reduction of nitrateand nitrite to hydroxylamine by the zinc amalgam reactor (Jones reductor)at pH 3.4 and reoxidation of the product with iron (III) in the presence offerrozine. Interference by high levels of nitrite could be eliminated with azidetreatment. Levels of nitrate of 0.1 mg/l could be detected with a precision of3% in the presence of large amounts of nitrite and chloride.

Nagashima et al. [154] used second derivative spectrometry to determinenitrate and nitrite in seawater. Samples of artifical seawater, pure water andchlorinatedwaterwerepumped throughacolumncontainingcadmium/copperpowder which converted nitrate to nitrite. The produced and existing nitritewas reduced to nitrogen monoxide, and the mixture was fed through a gas–liquid separator. Evolved nitrogen monoxide was purged by nitrogen intoa heated optical cell where the second derivative absorbance was recorded at214 nm. Various experimental conditions were investigated. A reducing agentof 0.13 M sodium iodide/13 M phosphoric acid, a mixing coil 5 m long operatedat 50 ◦C and a separator 50 cm long with the outer tube heated to 90 ◦C wereselected. Nitrite was determined directly, and nitrate indirectly, by measuringdifferences in nitrite with and without the converter.

The Department of the Environment UK [155] has described a number ofalternative methods for the determination of total oxidised nitrogen (nitrateand nitrite) in aqueous solution, while specific methods for nitrate and ni-trite are also included. Among the methods for total oxidised nitrogen, oneis based on the use of Devarda’s alloy for reduction of nitrate to ammonia,and another uses copperised cadmium wire for reducing nitrate to nitrite,which is determined spectrophotometrically. Nitrate may also be determinedspectrophotometrically after complex formation with sulfosalicylic acid; orfollowing reduction to ammonia, the ammonia is eliminated by distillationand determined titrimetrically. Other methods include direct nitrate determi-nation by ultraviolet spectrophotometry, measurements being made at 210 nm,and the use of a nitrate-selective electrode. Details of the scope, limits of de-tection, and preferred applications of the methods are given in each case.

Various other workers have discussed spectrophotometric methods for thedetermination of nitrate and nitrite [156–163].


Flow injection analysis has been adapted to automatic air-segmented contin-uous flow systems, e.g., the Technicon AutoAnalyzer system. Several reducing


agents such as zinc [164], cadmium [161], and amalgamated [163,165] or cop-perised cadmium [147, 148] have been investigated. Reductor columns are,however, difficult to operate in an air-segmented stream. This problem can beavoided by using the continuous flow injection technique developed by Ruzickaand Hansen [128, 166].

The flow injection technique is based on three main principles: sample in-jection, reproducible timing, and controlled dispersion [128]. The dispersioncan be described as limited, medium, or large; in a colorimetric system basedon a reaction between the sample and a suitable reagent, a medium dispersionis preferred. Thus in the flow injection determination of nitrate, the reductorcolumn should not excessively increase the dispersion. In a copperised cad-mium reductor, more than 90% of the total nitrate is reduced within 1–2 s withminimum risk of further reduction of nitrite [167]. Consequently, the reductorcan be made very small, which results in a minimal increase of dispersion.

In this development of a flow injection method for the determination of ni-trateandnitrite,Anderson[168]chose theShinn[155]methodtoreducenitrateand nitrite because of its high sensitivity and relative freedom from interfer-ences. Anderson [168] used flow injection in the photometric determinationof nitrite and nitrate with sulfanilamide and N-(1-naphthyl) ethylenediamineas reagents, as discussed next. The detection limit is 0.05 µm for nitrite and0.1 µm for nitrate at a total sample volume of 200 µL. Up to 30 samples can beanalysed per hour with relative precision of about 1%.

Figure 2.4. Flow diagram for the colorimetric determination of nitrite and nitrate. Theinternal diameter of the Tygon tubing was 0.64 mm for the 0.23 ml/min flow rates and1.30 mm for the 1.00 ml/min flow rates; all other tubing was 0.7 mm id. The reagents were(A) carrier stream, (B) sulfanilamide solution, and (C) N(1-naphthyl) ethylene diaminesolution. S denotes the point of injection and W waste. From [168]

2.20 Nitrate and Nitrite 75

A schematic diagram of the flow injection system used by Anderson [126] isshown in Fig. 2.4. An Ismatec model MP 13 peristaltic pump was used. Differentflow rates were obtained by changing the pump tube diameter, as indicatedin the legend to Fig. 2.4. The injection port was a rotary valve [131, 170]. Thesample volume could be varied between 10 and 1000 µl simply by changing thelength of the sample loop.

In the analysis of seawater, the only significant interference arises fromturbidity caused by particles in the sample. Prior filtration of the sample istherefore necessary. For anoxic waters, however, sulfide concentrations over2 µm were found to decrease the absorbance. This was overcome by adding anexcess of either Cd2+ or Hg2+ to the sample [171, 172].

A submersible flow injection-based sensor has been used to determinenitrite and nitrate in seawater. Detection limits of 0.1 µm for nitrate and a linearrange of 0.1–0.55 µM were achieved [173].

2.20.3Continuous Flow Analysis

Workers at the Department of the Environment UK [174] have describedcontinuous flow methods for the determination of total oxidised nitrogen andnitrite in seawater. Limits of detection are 1.3 µg/l (total oxidised nitrogen) and0.26 µg/l (nitrite). Within-batch standard deviations for total oxidised nitrogenrange from 0.28 µg/l to 17.5 µg/l at the total oxidised nitrogen level, to 0.96 µg/lat the 560 µg/l total oxidised nitrogen level. Within-batch standard deviationsfor nitrite range from 0.056 µg/l at the 3.5 µg/l nitrite level to 0.042 µg/l at the70 µg/l nitrite level.

Reduction of nitrate is achieved by the use of a mixing coil containingcopperised cadmium wire. Spectrophotometric evaluation of the nitrite pro-duced is achieved by the use of sulfanilamide N(1 naphthyl) ethylene diaminehydrochloride system.

2.20.4Reverse Phase Ion Interaction Liquid Chromatography

Ito et al. [175] used this technique employing octadecyl silane reverse phasecolumns coated with cetyltrimethyl ammonium chloride for the determinationof nitrite and nitrate in seawater.

2.20.5Miscellaneous

Sakamoto et al. [143] described an automated, near real-time analysis withmicroprocessor-controlled syringe pump modules for the determination of


nitrate in seawater. This provides an automated nitrate mapping system foruse on open ocean and coastal waters.

Sample preservation by pasteurisation has been applied to the preservationof nitrite and nitrate in seawater [152]. Samples can be stored for severalmonths before analysis.

2.21Perrhenate

Matthews and Riley [176] have described the following procedure for deter-mining down to 0.06 µg/l perrhenate in sea water. From 6 to 8 µg/l rheniumwas found in Atlantic sea water. The rhenium in a 15 litre sample of sea wateracidified with hydrochloric acid is concentrated by adsorption on a column ofDe-Acidite FF anion-exchange resin (Cl form), followed by elution with 4 mol/lnitric acid and evaporation of the eluate. The residue (0.2 ml), together withstandard and blanks, are irradiated in a thermal neutron flux of at least 3 ×1012

neutronscm–1 forat least 50h.After adecayperiodof twodays, the sample solu-tion and blank are treated with potassium perrhenate as carrier and evaporatedto dryness with a slight excess of sodium hydroxide. Each residue is dissolvedin 5 mol/l sodium hydroxide. Hydroxylammonium chloride is added (to reduceTc(VIII) which arises as 99mTc from activation of molybdenum present in thesamples), and the Re(VII) is extracted selectively with ethyl methyl ketone. Theextracts are evaporated, the residue is dissolved in formic acid:hydrochloricacid (19:1), the rhenium is adsorbed on a column of Dowex-1, and the col-umn is washed with the same acid mixture followed by water and 0.5 mol/lhydrochloric acid. The rhenium is eluted at 0 ◦C with acetone:hydrochloricacid (19:1), and is finally isolated by precipitation as tetraphenylarsonium per-rhenate. The precipitate is weighed to determine the chemical yield, and the186-rhenium activity is counted with an end-window Geiger–Muller tube. Theirradiated standards are dissolved in water together with potassium perrhen-ate. At a level of 0.057 µg/l rhenium, the coefficient of variation was ±7%.

2.22Phosphate

2.22.1Reverse Flow Injection Analysis

This technique differs from flow injection analysis in the sense that whereas inthe latter technique the sample plug is injected into a flowing stream of reagent,in the former technique plugs of reagent are injected into a continuous streamof the sample. Under these conditions the amount of sample in the zone of thereagent will increase as the dispersion increases. The sample will become well

2.22 Phosphate 77

Figure 2.5. Analytical manifolds for the determination of phosphate by flow injection anal-ysis (a) and reverse flow injection (b). The symbols S, M, and A are the seawater, mixedreagent, and the ascorbic acid solutions. The pump injection valve and detector are repre-sented by P, I, and D, respectively. W = waste. From [177]

mixedwith the reagentas its concentration increases, so thatwell-formedpeakswill result. Thus, an analysis can be successfully performed by this techniquewith a sample concentration in the zone of the reagent that is typically in therange from 67% to 90% of the sample concentration in the carrier stream.

Johnson and Petty [177] adapted reverse flow injection analysis to the well-known Murphy and Riley [178] colorimetric phosphomolybdate reductionmethod for the determination of phosphate.

Figure 2.5a and b show flow sheets for the determination of phosphate byflow injection analysis and reversed flow injection analysis, respectively.

The effects of residence time on the peak height of a 2.5 µmol/l phosphatestandard are shown in Fig. 2.6. The residence time was defined as the time frominjection of the reagents to the appearance of the maximum signal at the de-tector. Residence times were varied by changing the flow rate (1.0–3.5 ml/minin 0.5 ml/min increments) and the length of the reaction tube (0.5, 1.0, and1.5 m). The peak height increased rapidly with residence time in both 0.5 and0.8 mm id tubes, and then levelled off after 15 sec. The change is about thesame when the residence time is increased by using longer tubes. This suggeststhe increase peak height is due mainly to an increase in the extent of reaction.


Spencer and Brewer [111] have reviewed methods for the determination ofphosphate in seawater. Earlier methods for the determination of phosphatein seawater are subject to interferences, particularly by nitrate. In one early


Figure2.6. Effect of residence time on peakheight. The squares (0.5 mm id tubing) andcircles (0.8 mmid)wereobtainedbychang-ing the flow rate through a 1 m reactiontube. The square at the extreme right corre-sponds to 0.7 mm/min. The triangles wereobtained by using a 0.8 mm id tube anda flow rate of 2.0 ml/min. Residence timewas varied by changing the tube length.From [177]

method [179], the filtered seawater sample is acidified with nitric acid and per-chloric acid in the presence of hydroxylammonium chloride and ammoniumchloride, and the phosphomolybdate complex extracted with methylisobutylketone. This extract is then reduced with acidic stannous chloride and ascorbicacid, and the extinction measured at 725 nm. Hosokawa and Ohshima [180]heated the sample (50 ml) for 20 min on a boiling water bath with 2 ml Mo(VI)–Mo(V) reagent (hydrochloric acid added (10 ml) to Mo(VI) (2 M; 10 ml), thenzinc (0.3 g) added, and after the zinc has dissolved, hydrochloric acid wasadded to 100 ml). The mixture was cooled and the extinction measured at830 nm. The resulting blue complex remains stable for at least two months andhas an ε value of 26 000. Nitrate interferes if present in concentrations greaterthan 1 µg/l, and should first be reduced to nitrate with zinc in acidic medium.AsO3–

4 interferes at levels of 10 mg/l. The salt error was about 5% at a chlorideconcentration of 1.9%.

Isaeva [181] described a phosphomolybdate method for the determinationof phosphate in turbid seawater. Molybdenum titration methods are subject toextensive interferences and are not considered to be reliable when comparedwith more recently developed methods based on solvent extraction [182–187],such as solvent-extraction spectrophotometric determination of phosphate us-ing molybdate and malachite green [188]. In this method the ion pair formedbetween malachite green and phosphomolybdate is extracted from the seawa-ter sample with an organic solvent. This extraction achieves a useful 20-foldincrease in the concentration of the phosphate in the extract. The detectionlimit is about 0.1 µg/l, standard deviation 0.05 ng–1 (4.3 µg/l in tap water), andrelative standard deviation 1.1%. Most cations and anions found in non-salinewaters do not interfere, but arsenic (V) causes large positive errors.

2.22 Phosphate 79

In earlier work, Motomizu et al. [187] used ethyl violet as counter-ionfor phosphomolybdate. The spectrophotometric determination of phosphateby solvent extraction of molybdophosphate with ethyl violet is more sensitivethanpreviously reported [179–190], and less troublesome; a single extraction isadequate. However, in the determination of phosphorus at ng/l levels in watersit has certain disadvantages. First, the absorbance of the reagent blank becomestoo large for the concentration effect achieved by the solvent extraction to beof much use. For example, when 20 ml of sample solution and 5 ml of organicsolvent were used, the absorbance of the reagent blank was 0.14. Secondly, theshaking time needed was long and the colour of the extract faded gradually ifthe shaking lasted more than 30 min.

However, malachite green [188] gave a stable dark yellow species in 1.5 Msulfuric acid (probably a protonated one), whereas ethyl violet became colour-less within 30 min even in only 0.5 M sulfuric acid.

In the malachite green procedure, 10 ml of the sample solution containingup to 0.7 µg phosphorus as orthophosphate was transferred into a 25 ml testtube. To this solution was added 1 ml each of 4.5 M sulfuric acid and the reagentsolution. The solution was shaken with 5 ml of a 1:3 v/v mixture of toluene and4-methylpentan-2-one for 5 min. After phase separation, the absorbance of theorganic phase was measured at 630 nm against a reagent blank in 1 cm cells.

The absorption spectra of the ion pair shows a maximum at 632 nm. Lin-ear calibration graphs are obtained even when the aqueous phase volume isincreased to 100 ml (and 7 ml of extracting solvent are used). When 50 ml ofsample are used, 0.1 µg/l of phosphorus can be detected.

Arsenic (V) causes large positive errors – arsenic (V) at a concentration of10 µg/l produces an absorbance of 0.07, but can be masked with tartaric acid(added in thereagent solution).Whenarsenic (V)waspresentat concentrationsof 50 µg/l it was masked with 0.1 ml of 1 ×10–4 M sodium thiosulfate addedafter the sulfuric acid.

For a 25 ml sample, almost all of the foreign ions tested, when present atthe concentrations listed in Table 2.4, produced an absorbance of less than0.005. The recovery of phosphorus was good: 99–103%. The relative standarddeviation for phosphorus was 0.6% for 21.0 µg/l in seawater.

A commonly used procedure for the determination of phosphate in seawa-ter and estuarine waters uses the formation of the molybdenum blue complexat 35–40 ◦C in an autoanalyser and spectrophotometric evaluation of the re-sulting colour. Unfortunately, when applied to seawater samples, dependingon the chloride content of the sample, peak distortion or even negative peaksoccur which make it impossible to obtain reliable phosphate values (Fig. 2.7).This effect can be overcome by the replacement of the distilled water-washsolution used in such methods by a solution of sodium chloride of an appro-priate concentration related to the chloride concentration of the sample. Thechloride content of the wash solution need not be exactly equal to that of thesample. For chloride contents in the sample up to 18 000 mg/l (i.e., seawater),


Table 2.4. Effect of other ions [188]

Ion Added as Concentration mol/l Absorbance∗

None 0.549Na+, Cl– NaCl 0.3 0.545K+ KCl 1 ×10–2 0.548SO2–

4 Na2SO4 1 ×10–2 0.552Ca2+ CaCl2 1 ×10–2 0.551Mg2+ MgCl2 1 ×10–2 0.548HCO–

3 NaHCO3 1 ×10–2 0.556Al+3 KAl(SO4)2 1 ×10–3 0.550Fe+

3 FeNH4(SO4)2 1 ×10–3 0.551NH+

4 NH4Cl 1 ×10–3 0.557NO–

3 NaNO3 1 ×10–3 0.548Cr3+ CrCl3 1 ×10–3 0.549Mn2+ MnCl2 1 ×10–3 0.549Co2+ CoCl2 1 ×10–3 0.549Cd2+ CdCl2 1 ×10–3 0.552Ni2+ NiCl2 1 ×10–3 0.549Cu2+ CuCl2 1 ×10–3 0.553Zn2+ ZnCl2 1 ×10–3 0.546Br– NaBr 1 ×10–3 0.546B(III) Na2B4O7 1 ×10–3 0.552SiO2–

3 Na2SiO3 5 ×10–4 0.553VO3 NH4VO3 1 ×10–5 0.556WO2–

4 (NH4)10W12O41 1 ×10–5 0.554I– NaI 1 ×10–5 0.550Laurylsulfate Na salt 5 ×10–6 0.549Laurylbezenesulfonate Na salt 5 ×10–6 0.547ClO–

4 NaClO4 1 ×10–6 0.557Ti+4 TiCl4 1 ×10–6 0.551

∗ Sample solution 10 cm3 organic phase 5 cm3; phosphorus 0.366 µg;measured against reagent blank. †Tolerance limit

the chloride concentration in the wash should be within ±15% of that in thesample. The use of saline standards is optional, but the use of saline controlsolutions is mandatory. Using good equipment, down to 0.02 mg/l phosphatecan be determined by such procedures. For chloride contents above 18 000 mg/lthe chloride content of the wash should be within ±5% of that in the sample.

Flynn and Meehan [191] have described a solvent extraction phospho-molybdate method using isoamyl alcohol for monitoring the concentration of32P in sea and coastal waters near nuclear generating stations.

Deans [192] have proposed a method for the colorimetric determination oftraces of phosphorus with molybdenum blue, making use of the laser-inducedthermal lensing effect. The procedure is described, and the results obtained onsamples of sea water and lake water are presented.

2.22 Phosphate 81

Figure 2.7. Interferenceby chloride in the auto-analyser determinationof phosphate

In this method an Ar+ laser-pumped rhodamine 101 laser was used asthe heat and probe source. The signal from a silicon photocell with a 1 mm2

photosensitive surface, which was used as a laser radiation detector, was pro-cessed with an inexpensive personal computer. The detector limit is 5 pg ofphosphorus/ml.

Yoshimura et al. [193] carried out microdeterminations of phosphate bygel-phase colorimetry with molybdenum blue. In this method phosphate re-acted with molybdate in acidic conditions to produce 12-phosphomolybdate.The blue species of phosphomolybdate were reduced by ascorbic acid in thepresence of antimonyl ions and adsorbed on to Sephadex G-25 gel beads. Atten-uation at 836 and 416 nm (adsorption maximum and minimum wavelengths)was measured, and the difference was used to determine trace levels of phos-phate. The effect of nitrate, sulfate, silicic acid, arsenate, aluminium, titanium,iron, manganese, copper, and humic acid on the determination were examined.

Eberlein and Kattner [194] described an automated method for the deter-mination of orthophosphate and total dissolved phosphorus in the marineenvironment. Separate aliquots of filtered seawater samples were used for thedetermination orthophosphate and total dissolved phosphorus in the concen-tration range 0.01–5 µg/l phosphorus. The digestion mixture for total dissolvedphosphorus consisted of sodium hydroxide (1.5 g), potassium peroxidisulfate(5 g) and boric acid (3 g) dissolved in doubly distilled water (100 ml). Seawatersamples (50 ml) were mixed with the digestion reagent, heated under pressureat 115–120 ◦C for 2 h, cooled, and stored before determination in the auto-analyser system. For total phosphorus, extra ascorbic acid was added to theaerosol water of the autoanalyser manifold before the reagents used for themolybdenum blue reaction were added. For measurement of orthophosphate,a phosphate working reagent composed of sulfuric acid, ammonium molyb-


date, ascorbic acid, and antimony potassium tartrate solution was prepared.Up to 30 samples per hour could be analysed.

Airey et al. [195] have described a method for the removal of sulfide priorto the determination of phosphate in anoxic estuarine waters. Mercury (II)chloride was used to precipitate free sulfide from samples of anoxic water. Thesulfite-free supernatant liquid was used to estimate sulfide by measuring theconcentration of unreacted mercury (II), as well as to determine phosphateby the spectrophotometric method in which sulfide interferes. The detectionlimit for phosphate was 1 µg/l.


Tyree and Bynum [132] have described an ion chromatographic method forthe determination of phosphate and nitrate in seawater.

2.23Propionate


See Sect. 2.1.1.

2.24Pyruvate


See Sect. 2.1.1.

2.25Selenate/Selenite

2.25.1Fluorometric Method

Itoh et al. [196] determined selenium (IV) in sea and estuarine waters by ananion-exchange resin modified with bismuthiol (II) and diaminonaphthalenefluorophotometry. An Amberlite IRA-400 anion exchange resin was modifiedby mixing with an aqueous solution of bismuthiol (II) to give 0.2 mmol per g of

2.26 Silicate 83

resin. Sample solution was eluted through a column packed with the modifiedresin.

Selenium (IV) adsorbed as selenotrisulfate was then eluted from the columnwith either 0.1 M penicillamine or 0.1 M cysteine. The eluate was then subjectedto an acid digestion procedure to reduce selenium to the tetravalent state withdiaminonaphthalene for fluorometric determination. Approximate agreementwith the tellurium coprecipitation method was obtained. The application ofboth methods to the analysis of estuarine waters permitted the separate de-termination of both selenium (IV) and selenium (VI), since the telluriumcoprecipitation methods did not differentiate between the two species.

2.26Silicate


Silicon is an essential and, in some cases, a growth-limiting micronutrientfor marine organisms that form siliceous skeletons [197]. Concentrations ofavailable silicon are often low (< 1 µM) in surface waters of high productivity.Although most of the silicon is recycled near the surface, a fraction reachesthe sediment surface, where it may dissolve. As a result, deep waters mayhave dissolved silicon concentration as high as 200 µM. Upwelling returns thedissolved silicon to the surface, and the cycle is repeated. Dissolved silicon is animportant species to monitor because of its influence on plankton growth [197]and because of its use as a tracer of water mass movement [198].

Various approaches to the analysis of dissolved silicon have been tried. Mostof themarebasedon the formationofβ-molybdosilic acid [199–203].Dissolvedsilicon exists in seawater almost entirely as undissociated orthosilicic acid. Thisform and its dimer, termed “reactive silicate”, combine with molybdosilicicacid to form α- and β-molybdosilicic acid [180]. The molybdosilicic acid canbe reduced to molybdenum blue, which is determined photometrically [206].The photometric determination of silicate as molybdenum blue is sufficientlysensitive for most seawater samples. It is amenable to automated analysisby segmented continuous flow analysers [206–208]. Most recent analyses ofsilicate in seawater have, therefore, used this chemistry. Furthermore, reactivesilicate is probably the only silicon species in seawater that can be used bysiliceous organisms [204].

Spencer and Brewer [111] have reviewed methods for the determinationof silicate in seawater. Various workers [209–212] have studied the applica-tion of molybdosilicate spectrophotometric methods to the determination ofsilicate in seawater. In general, these methods give anomalous results due, itis believed, to erratic blanks and uncertainty regarding the structure of thesilicomolybdate formed.


Brzelinski and Nelson [214] have described a solvent extraction procedurefor the spectrophotometric determination of nanomolar concentrations ofsilicic acid in seawater.

β-silicomolybdic acid was formed by the reaction of silicic and molybdicacids at low pH. The combined acid was extracted into n-butanol and reducedwith a mixture of p-methylaminophenol sulfate and sulfite. After phase sep-aration, the samples were cleaned by chilling in ice followed by centrifuging,and the colour was evaluated spectrophotometrically at 810 nm. The molarabsorbance of the mixed acid was 229 000 in seawater and the precision was±2.5 nmol/l silicon for concentrations less than or equal to standard aqueousanalyses. Sensitivity in seawater was 70% of that in distilled water because ofa significant salt effect. Natural concentrations of arsenate, arsenite, and ger-manic acid caused negligible interference. Phosphate interference was equiv-alent to 10–12 nmol/l silicon over a broad range of phosphate concentrations.


Flow injection analysis is a rapid method of automated chemical analysis thatallows for quasi-continuous recording of nutrient concentrations in a flowingstream of seawater. The apparatus used for flow injection analysis is generallyless expensive and more rugged than that used in segmented continuous flowanalysis. A modified flow injection analysis procedure, called reverse flowinjection analysis, was adopted by Thompson et al. [213] and has been adaptedfor the analysis of dissolved silicate in seawater. The reagent is injected intothe sample stream in reverse flow injection analysis, rather than vice versa asin flow injection analysis. This results in an increase in sensitivity.

This analytical procedure is based on an optimum analysis condition forsegmented continuous flow analysis. The sample is combined with a molyb-date solution at a pH between 1.4 and 1.8 to form the β-molybdosilicic acid.After an appropriate time for reaction, a solution of oxalic acid is added, whichtransforms the excess molybdate to a non-reducible form. The oxalic acid alsosuppresses the interference from phosphate by decomposing phosphomolyb-dic acid. Finally, a reductant is added to form molybdenum blue. Both ascorbicacid and stannous chloride were tested as reductants.

In the determination of silicate, a detection limit of 0.5 µmol/l was achievedwith a relative precision of better than 1% at concentrations above 10 µmolessilicon/l.

2.26.3Ion Exclusion Chromatography

Hioki et al. [215] have described an on-line determination of dissolved silicain seawater by ion exclusion chromatography in combination with inductivelycoupled plasma emission spectrometry.

2.27 Sulfide 85

This method was developed as a second independent method to comple-ment the usual colorimetric procedure in the determination of a certifiedconcentration of dissolved silica in a seawater reference material. Ion exclu-sion affords a separation of the dissolved silica not only from major seawatercations but also from potentially interfering anions. The detection unit limit,conservatively estimated at 2.3 ng/g Si (0.08 µm), is superior to that achievableby direct analysis using inductively coupled plasma emission spectrometry.

Other studies on the application of this technique have been carried out[216–219].

2.27Sulfide

2.27.1Gas Chromatography

Cutter and Oatts [220] determined dissolved sulfide and studied sedimen-tary sulfur speciation using gas chromatography in conjunction with a pho-toionisation detector. The determination of dissolved sulfide and sedimen-tary sulfur is important to studies of trace element cycling in the aquaticenvironment. A method employing selective generation of hydrogen sulfide,liquid-nitrogen-cooled trapping, and subsequent gas chromatographic separa-tion/photoionisation detection, has been developed for such studies. Dissolvedsulfide is determined via acidification and gas stripping of a water sample, witha detection limit of 12.7 nM and a precision of 1% (relative standard devia-tion). With preconcentration steps the detection limit is 0.13 nM. The detectionlimit for those sulfur species is 6.1 µg of S/g, with the precision not exceeding7% (relative standard deviation). This method is rapid and free of chemicalinterferences; field determinations are possible.

Leck and Bagander [221] determined reduced sulfide compounds (hydro-gen sulfide, methyl mercaptan, carbon disulfide, dimethyl sulfide, and dimeth-yldisulfide) in water by gas chromatography using flame detection. Detectionlimits ranged from 0.2 ng/l for carbon disulfide to 0.6 ng/l for methyl mercap-tan. Hydrogen sulfide was determined at the 1 ng/l level.

2.27.2Capillary Isotachoelectrophoresis

Fukishi and Hiiro [222] determined sulfide in seawater by this technique. Themethod is based on the generation of hydrogen sulfide by the addition ofsulfuric acid to the water sample. The gas permeated through a microporouspolytetrafluoroethylene (PTFE) tube, and was collected in a sodium hydroxidesolution. The carbon dioxide in the permeate was removed by adding a bariumcation-exchange resin to the sodium hydroxide solution. Injection into the


isotachophoresis apparatus followed. A linear graph for sulfide concentrationsof up to 2 mg/l was obtained.

2.28Sulfate


Sulphate has been determined in seawater by photometric titration with hydro-chloric acid in dimethyl sulfoxide [223]. The sample (5 ml) is slowly added todimethyl sulfoxide (230 ml) and titrated with 0.02 M hydrochloric acid (stan-dardised against sulfate) with bromo-cresol green as indicator. Since borate,carbonate, and bicarbonate interfere, a separate determination of alkalinity isnecessary.

Lebel [224] has described an automated chelometric method for the de-termination of sulfate in seawater. This method utilises the potentiometricend-point method for back titration of excess barium against EDTA followingprecipitation of sulfate as barium sulfate. An amalgamated silver electrodewas used in conjunction with a calomel reference electrode in an automatictitration assembly consisting of a 2.5 ml autoburette and a pH meter coupled toa recorder. Recovery of added sulfate was between 99 and 101%, and standarddeviations of successive analyses were less than 0.5 of the mean.

2.28.2Inductively Coupled Plasma Atomic Emission Spectrometry

Inductively coupled plasma atomic emission spectrometry has also been usedto determine sulfate directly in non-saline waters [225].

By monitoring the 180.73 nm sulfur line, it was possible to detect 0.08 mg/lsulfate with a precision of 0.8% relative standard deviation at the 200 mg/lsulfate level. Instrument operating conditions are shown in Table 2.5.

Of the elements characteristically present as major components of non-saline waters, only calcium produced a slight interference. A weak calciumline at approximately 190.73 nm partially overlaps the sulfur line such that1000 mg/l Ca produces an apparent signal of 25 mg/l sulfate. Correction for thisinterference is easily achieved by establishing the relationship between calciumconcentration and apparent sulfate signal, and inserting this information in thecontrolling software. The effect of calcium was then automatically subtractedduring the measurement of the sample.

Non-saline water samples are normally acidified to stabilise them duringstorage. There is an effect due to hydrochloric acid concentration on the sul-fur emission signal. This effect is conveniently overcome by making the acidcontent of samples and standards identical at, for example, 1 vol%.

2.28 Sulfate 87

Table 2.5. Plasma system components, and operating conditions [225]

Spectrometer ARL 3400 C Quantovac; 1 m Paschen-Range monuing; grating ruled1080 lines mm–1, 0.309 nm mm–1 reciprocal linear dispersion (3rdorder). Primary slit width 20 µm, secondary slit 50 µm, HamamatsuR-306 photomultiplier tube

Wavelength 180.73 nm, 3rd orderReadout Digital readout of mean signal from three – 10 s integrationsFrequency 27.12 MHzForward power 1080 WReflected power 5 WObservation height 17 mm above load coilArgon flow rates Coolant 11 L/min, plasma 1.2 L/min, carrier 1 L/min, light path purge

1.5 L/minNebuliser Concentric glass, JE Meinhard model TR-30-A3

The accuracy of the inductively-coupled plasma procedure was assessed byanalysing waters of known sulfate composition, and by comparing measuredsulfate values for a wide range of samples with those obtained for the samewaters by an automated spectrophotometric procedure. Good agreement isobtained between the derived sulfate measurements and the normal values forInternational Standard Sea Water and an EPA Quality Control Standard.

2.28.3Polarography

Polarography has also been used to determine sulfate [226]. In this methodsulfate is precipitated as lead sulfate in a 20% alcoholic medium, followed bypolarographic determination of the excess lead. It is claimed that there is littleinterference in this method; 25–30 samples can be analysed in a five hourperiod.

In this procedure up to 70 ml seawater, 10 ml 0.01 M lead nitrate, 20 ml 95%ethanol, 0.2 ml of 0.1% methyl red maximum current suppressor, and twodrops of 3 M nitric acid are added to a 100 ml volumetric flask, the contents ofwhich are then diluted to 100 ml with distilled water.

The dropping mercury electrode had a drop time of 3–4 s under an openhead of 50 cm Hg. A saturated calomel electrode was the reference electrode.Before recording, the solutions were shaken well. After recording, the elec-trodes were well rinsed with distilled water and wiped dry. A starting potentialof 0.2 V was used and the solutions were degassed with dry nitrogen for 2 minprior to recording. A full-scale sensitivity of 10 µA was used.

As an indication of the accuracy of the technique the correlation coefficientsfor the calibration curves indicate a precision of 0.2% or better. The standarddeviation of the measurement samples is 0.70%.


The precision of the technique indicates its suitability for sulfate analysis.The technique appears not to be subject to any interference normally presentin seawater samples.


Singh et al. [227] have determined sulfate in deep sub-surface waters by sup-pressed ion chromatography.

2.29Valerate


See Sect. 2.1.1.

2.30Multianion Analysis

2.30.1Spectrophotometric Methods, Phosphate, Arsenate, Arsenite, and Sulfide

NasuandKan[228]determinedphosphate, arsenate, andarsenite innon-salinewater by flotation spectrophotometry and extraction indirect atomic absorp-tion spectrometry, using malachite green as an ion-pair reagent. A floated(between aqueous and organic phase) ion pair of malachite green with molyb-dophosphate was dissolved by the addition of methanol to the organic layer.Phosphate and arsenate were determined by measuring absorbance of theorganic phase. An oxidative (potassium dichromate) or a reductive (sodiumthiosulfate) reaction of arsenic was used for the determination of phosphate,arsenate, and arsenite. A positive interference effect was observed in the pres-ence of large amounts of silicon. This was overcome by acidification withconcentrated hydrochloric acid. The method was applied to samples of hotspring water, sea water, and ground water, with almost complete recovery ofadded amounts (0.04–0.2 µg/ml).

Johnson and Pilson [229] have described a spectrophotometric molybde-num blue method for the determination of phosphate, arsenate, and arsenite inestuary water and sea water. A reducing reagent is used to lower the oxidationstate of any arsenic present to +3, which eliminates any absorbance caused bymolybdoarsenate, since arsenite will not form the molybdenum complex. Thisresults in an absorbance value for phosphate only.

2.30 Multianion Analysis 89

A commonly used procedure for the determination of phosphate in seawater and estuarine waters involves the formation of the molybdenum bluecomplex at 35–40 ◦C in an autoanalyser followed by spectrophotometric evalu-ation of the resulting colour. Unfortunately when applied to sea water samples,depending on the chloride content of the sample, peak distortion or even nega-tive peaks occur, which make it impossible to obtain reliable phosphate values.This effect can be overcome by the replacement of the distilled water used insuch methods by a solution of sodium chloride of an appropriate concentrationrelated to the chloride concentration of the sample (see Sect. 2.22.2).

Airey et al. [230] have described a method for the removal of sulfide priorto the determination of phosphate in anoxic estuarine waters. Mercury (II)chloride was used to precipitate free sulfide from samples of anoxic water. Thesulfide-free supernatant liquid was used to estimate sulfide by measuring theconcentration of unreacted mercury (II), as well as to determine phosphate bya spectrophotometric method in which sulfide interferes. The detection limitfor phosphate was 1 µg/l.

2.30.2Electrostatic Ion Chromatography, Bromide, Nitrate, and Iodide

In a series of methods [231–237] employing stationary phases coated withzwitterionic surfactants, (i.e., those containing both positively and negativelycharged functional groups but carrying no formal net charge), it has beendemonstrated that inorganic anions can be separated using pure water aseluent, with unique separation selectivity. This method has been termed elec-trostatic chromatography: the separation has been attributed to a simultane-ous electrostatic attraction and repulsion mechanism occurring at both thepositive and the negative charges on the stationary phase. A drawback of elec-trostatic ion chromatography is that when the sample contains multiple anionsand cations, each analyte anion may be eluted as more than one peak, witheach peak being a specific combination of the anion with one of the cations ofthe sample. Recently, Hu and Haddad [238] showed that addition of a smallquantity of a suitable electrolyte to the eluent causes analyte anions to beeluted only as a single peak, irrespective of the number and type of cationsin the sample. In further studies, Hu et al. [239] investigated the separationmechanism in more detail and have applied the method to the determinationof nitrate, bromide, and iodide in seawater.

The determination of nitrate, bromide, and iodide in seawater is of impor-tance to oceanographic research [240–242]. Ion chromatography is generallyinapplicable to this analysis for several reasons. First, the large number of ma-trix ions (chloride, sulfate) saturates the active sites of the stationary phase andthereby impedes the separation of the target analytes. Second, the high ionicstrength of the sample causes self-elution of the sample band during injection,leading to peak broadening and loss of separation efficiency. Third, the levels


of the target analytes are often very low, so that detection becomes a majorproblem, especially when the eluted peaks are poorly defined in shape.

To overcome these difficulties, Ito and Sunahara [102] suggested the useof the matrix ions as the eluent ions, for example the use of relatively highconcentrations of sodium chloride solution.

In the most recent method described by Hu et al. [239] for the direct de-termination of ultraviolet-absorbing inorganic anions in saline matrixes, anoctadecylsilica column modified with a zwitterionic surfactant [3-(N, N-di-methylmyristylammonio)propanesulfate] is used as the stationary phase, andan electrolytic solution is used as the eluent. Under these conditions, the matrixspecies (such as chloride and sulfate) are only retained weakly and show little orno interference. It is proposed that a binary electrical double layer is establishedby retention of the eluent cations on the negatively charged (sulfonate) func-tional groups of the zwitterionic surfactant, forming a cation-binary electricaldouble layer.

2.30.3Miscellaneous

Determinations of the so-called “micronutrients”, (e.g., nitrate, phosphate,and dissolved silica), in seawater are among the most commonly performedanalyses in oceanographic research and survey work. Surprisingly, there existsno certified reference material that can be used to check the accuracy of suchanalyses, although intercomparisonexerciseshavebeenconductedona regularbasis [243]. The lack of seawater certified reference material for micronutrientscan be attributed both to the difficulty of preparing a suitable material withan adequate shelf life [244, 245] and to the dearth of independent methodsavailable for the determinations.

The National Research Council of Canada has undertaken a project, incollaboration with Bedford Institute of Oceanography of the Canadian De-partment of Fisheries and Oceans, to address the need for a seawater certifiedreference material for micronutrients with the initial objective being the prepa-ration of a material with certified values for nitrate, phosphate, and dissolvedsilica.

2.31pH

Sedijl and Lu [246] have described a colorimetric method using optical deflec-tion for the determination of the pH of seawater.

Yaakov and Ruth [247, 248] have described an improved in situ pH sensorfor oceanographic and limnological applications. They report that an accuracyof 0.002 pH units can be achieved.

2.32 Suspended Solids 91

Khoo et al. [249] have reported on the measurement of standard potentials,hence hydrogen ion concentrations, in seawaters of salinities between 20 and45‰ at 5–40 ◦C.

2.32Suspended Solids

Suspended solids determinations on non-saline samples are a routine pro-cedure that does not present any problems. The solids, after filtration ona 0.45 nm glass fibre disc are dried at 105 ◦C to obtain the moisture contentand at 450–500 ◦C to obtain the organic content. However, the application ofthis procedure to saline samples does present some problems, owing to errorscaused by the occlusion of sea salts on the filter disc and the filtered solids.One innovation that has been adopted to correct for high solids contents onsuch samples has been to filter the sample through a double layer of glass fibrepaper. The weight increase of the lower paper is due to the occluded salts whilstthat on the top paper is due to salts plus sample solids. The corrected solidscontent is then obtained by subtracting the weight increase of the lower discfrom that of the upper disc. Unfortunately, results obtained by this modified

Table 2.6. Anion preconcentration (from author’s own files)

Anion Preconcentrationmethod

Analytical finish Detectionlimit (µg/l)

Section Reference

SbO4 Solvent extraction Spectrophotometric – 2.30.1 [228]AsO4AsO3CrO4 Fe(OH)3

coprecipitationAtomic absorptionspectrometry

0.04 2.10.1 [35–38]

Immobiliseddiphenyl carbazone

Atomic absorptionspectrometry

– 2.10.1 [250]

F Solvent extractionof ternarylanthanum alizarincomplexonefluoride

Atomic emissionspectrometry

0.6 2.11 [251]

MoO4 Adsorption ofTiron complex onanion exchangeresin

– – 2.16 [252]

Adsorption onbasic anionexchange resin

Atomic absorptionspectrometry

< 10 2.16 [253]

Perrhenate Adsorption onDe-Acidite anionexchange resin

Neutron activationanalysis

0.06 2.20 [176]


procedure are still unreliable, becoming more so at higher salt concentrationsin the sample.

A further simple innovation gives much more reliable solids results. In thisthe glass fibre disc mounted in its porcelain, glass, or plastic holder is firstwetted with distilled water, and then the sample filtered through and washedwith several small portions of distilled water without allowing the disc tobecome dry. It is believed that filling the air spaces in the annulus of the disctrapped in the holder with distilled water is the reason why better results areobtained by this procedure.

2.33Anion Preconcentration

Methods of improving the detection limits for anions by preconcentration arereviewed in Table 2.6.

References

1. Yang X-H, Lee C, Schanton MI (1993) Anal Chem 65:8572. Gran G (1952) Analyst 77:6613. Culberson C, Pytowicz RM, Hawley JE (1970) J Mar Res 28:154. Pearson F (1981) J Water Pollution Control 53:12435. Dickson AG (1981) Deep-Sea Res 28A:6096. Sass E, Ben-Yaakov S (1977) Mar Chem 5:1837. Bates RG (1964) Determination of pH – theory and practice. Wiley, New York8. Ben-Yaakov S, Lorch Y (1983) Mar Chem 13:2939. Ben-Yaakov S, Raviv R, Guterman H, Davan A, Lazar B (1982) Talanta 29:267

10. Keir RS, Kounaves SP, Sirino AL (1977) Anal Chim Acta 91:18111. Johansson A, Johnansson S, Gran G (1972) Analyst 108:108612. Stuart WA, Lister AR Report from the UK Atomic Energy Authority AERE-M2250-5PP,

AERE, Harwell, UK13. Hansson I, Jagner D (1973) Anal Chim Acta 65:36314. Haywood MG, Riley JP (1976) Anal Chim Acta 219:8515. Chaime F, Zeitlin H (1974) Separ Sci 1:916. Johnson O (1972) Anal Abst 2768:2217. Sato S (1983) Anal Chim Acta 151:46518. Macalady DL, Carpenter DH, Moore CA (1977) Science 195:133519. Carpenter JH (1965) Limnol Oceanogr 10:13520. Lewin M, Avrahani M (1950) J Am Chem Soc 47:41921. Kolthoff JM, Yutzy A (1937) Ind Eng Chem 9:7522. Rose O, Cuttita A (1968) Advanced X-Ray Analysis 11:2323. Basel CL, Defreeze JD, Whittemore DO (1982) Anal Chem 54:209024. Walters FH (1984) Anal Lett 17:168125. Sichère M-C, Cesbron F, Zuppi G-M (1978) Anal Chim Acta 28:29926. Fukushi K, Hiro K (1987) Bunseki Kagaku 36:71227. Jagner D, Aren K (1970) Anal Chim Acta 52:49128. Jagner D (1970) Anal Chim Acta 50:15

References 93

29. Grasshoff K, Wenck A (1972) Mercantile Journal Cons Perm International 522:3430. Noborn O (1972) Japan Analyst 780:2131. Meyer RE, Dosey FA, Lontz PH (1968) J Electroanal Chem 99:1932. Wilson TRS (1971) Limnol Oceanogr 16:58133. Batley GE, Matousek JP (1980) Anal Chem 52:157034. De Jong GJ, Brinkman UAT (1978) Anal Chim Acta 98:24335. Cranston RE, Murray JW (1978) Anal Chim Acta 99:27536. Cranston RE, Murray JW (1980) Limnol Oceanogr 25:110437. Mullins TL (1984) Anal Chim Acta 165:9738. Nakayama E, Toroko H, Fujinaga T (1981) Anal Chim Acta 131:24739. Nakayama E, Kuwamoto H, Tokoro H, Tsurubo S, Fujinaga T (1981) Anal Chim Acta

130:28940. Grimaud D, Michard G (1974) Mar Chem 2:22941. Nakayama E, Kuwamoto T, Tokoro H, Tsurubo S, Fujinaga T (1981) Anal Chim Acta

130:29042. Nakayama E, Kuwamoto T, Tsurubo S, Fujinaka T Anal Chim Acta (1981) 130:40143. Nakayama E, Kuwamoto T, Tsurubo S, Fujinaka T Anal Chim Acta (1981) 130:40244. Ishibashi M, Shigematsu T (1950) Bull Inst Chem Res Kyoto University 23:5945. Chuecas L, Riley JP (1966) Anal Chim Acta 35:24046. Kuwamoto T, Murai S (1970) Preliminary Report of the Hakuho-maru Cruise KH-68-4,

Ocean Research Institute University, Tokyo, 7247. Fukai R (1967) Nature 213:90148. Fukai R, Vas D (1967) Journal Oceanographic Society, Japan 23:29849. Yamamoto T, Kadowski S, Carpenter JH (1974) Geochem J 8:12350. Ahern F, Eckert JM, Payne NC, Williams KL (1985) Anal Chim Acta 175:14751. Fletsch RA, Richards FA (1970) Anal Chem 42:140052. Greenhalgh R, Riley JP (1961) Anal Chim Acta 25:67953. Brewer PG, Spencer DW, Wilkniss (1970) Deep-Sea Res 17:154. Warner TB (1969) Science 165:17855. Windom HI (1971) Limnol Oceanogr 16:80656. Orion, Research Analytical Methods Guide 6th Edition (1973)57. Anfalt T, Jagner D (1971) Anal Chim Acta 53:1358. Rix CJ, Bond AH, Smith JD (1976) Anal Chem 48:123659. Barendrecht E (1967) In: Electroanalytical Chemistry (ed AJ Bond) Volume 2, Marcel

Dekker, New York, p. 5360. Brand HJD, Rechnitz GA (1970) Anal Chem 42:117261. Baumann EW (1970) Anal Chim Acta 42:12762. Craggs A, Moody GJ, Thomas JDR (1974) J Chem Educ 81:54163. Liberti A, Mascini M (1969) Anal Chem 41:67664. Baumann EW (1971) Anal Chim Acta 54:18965. Frani MS, Ross JW (1966) Science 154:155366. Riley JP, Chester R (1971) In: Introduction to Marine Chemistry (eds Riley JP, Chester

R) Academic Press, London, p. 70–8067. Bond AM, Hefter G (1971) J Inorg Nucl Chem 33:42968. Greenhalgh R, Riley JP (1961) Anal Chim Acta 25:67969. Warner TB (1971) Deep-Sea Res 18:125570. Special Publication No17 (1964) The Chemical Society, UK71. Ke PJ, Regier LW (1971) Anal Chim Acta 53:2372. Ke PJ, Regier LW (1970) Anal Abst 19:239273. Wilkniss PK, Linnebon WJ (1968) Limnol Oceanogr 13:530


74. Gomez MG, Palacios Corvillo MA, Camara Rica C (1988) Analyst 113:110975. Kieber DG, Vaughan GM, Mopper K (1988) Anal Chem 60:165476. William PM, Robertson K (1980) Journal of Water Pollution Control Federation 52:216777. Truesdale VW (1978) Mar Chem 6:25378. Truesdale VW (1978) Mar Chem 6:15079. Truesdale VW (1974) Deep-Sea Res 21:76180. Truesdale VW, Smith JS (1979) Mar Chem 7:13381. Truesdale VW, Smith JS (1975) Analyst 100:11182. Herring JR, Liss PS (1974) Deep-Sea Res 21:77783. Schnepfe MM (1972) Anal Chim Acta 58:8384. Tsunogai S (1971) Deep-Sea Res 18:91385. Tsunogai S, Henmi T (1971) J Oceanogr (Japan) 27:6786. Wong GTF, Brewer PG (1976) Anal Chim Acta 81:8187. Smith JD, Butler ECV, Airey D, Sandars G (1990) Mar Chem 28:35388. Wong GTF, Zhang L (1992) Cont Shelf Res 12:71789. Nakayama E, Kimoto T, Okazaki S (1985) Anal Chem 57:116090. Nakayama E, Kimoto T, Isshiki K, Sohrin Y, Okazaki S (1989) Mar Chem 27:10591. Takayanagi K, Wong GTF (1986) Talanta 33:45192. Luther III GW, Swartz CB, Ullman WJ (1988) Anal Chem 60:172193. Jickells TD, Biyd SS, Knap AH (1988) Mar Chem 24:6194. Wong GTF, Takayanagi K, Todd JF (1985) Mar Chem 17:17795. McTaggart AR, Butler ECV, Haddad PR, Middleton JH (1994) Mar Chem 7:15996. Tsunogai S, Sase T (1969) Deep-Sea Res 10:48997. Crompton TR (1977) Chemical Analysis of Organo-metallic Compounds, Volume 5,

Academic Press, New York, p. 13298. Sugawara A (1956) Anal Abst 3:116999. Matthews AD, Riley JP (1970) Anal Chim Acta 51:295

100. Shizuo T (1971) Anal Chim Acta 55:444101. Butler ECV, Gershey RM (1984)Anal Chim Acta 164:153102. Ito K, Sunahara H (1988) Bunseki Kagaku 37:292103. Ito K (1997) Anal Chem 69:3029104. Buchberger W, Huemer P, Winsauer K(1985) Z Wasser Abwass Fors 18:194105. Krishnamoorthy KR, Iyer RK (1987) Proceedings of the Indian Academy of Science,

Chemical Sciences 98:177106. Varma KK, Jaia A, Varma A (1992) Anal Chem 64:1484107. Yoshida I, Taekshita R, Ueno K (1987) Kenkyu, Hokaku-Kuwamoto Kogyo Daigaku

12:159108. Hara H, Okazaki S (1984) Analyst 109:1317109. Van Staden JF (1985) Fresen Z Anal Chem 322:336110. Basson WP, Van Staden JF (1981) Water Res 15:333111. Spencer DW, Brewer PG (1970) Crit Rev Solid State 1:409112. Matsunaga K, Nishimura M (1974) Anal Chim Acta 73:204113. Matsunaga K, Nishimura M (1969) Anal Chim Acta 45:350114. Malhotra SK, Zanoni AD (1970) J Am Water Works Assoc 62:568115. Lambert RS, Du Bois RJ (1971) Anal Chem 43:955116. Joos MH (1984) Water Res 18:643117. Mubarek A, Howald RA, Woodruff IR (1977) Anal Chem 49:857118. Mertens J, Massart DL (1971) Bulletin of the Belgium Chemical Society 80:151119. Hoather RC (1953) Applications of Spectrophotometry for the Examination of Water

Proceedings of the Society of Water Treatment and Examination 2:9

References 95

120. Hoather RC, Rackman RF (1959) Analyst 84:548121. Goldman E, Jacobs RJ (1961) J Am Water Works Assoc 53:187122. Morries E (1971) Proceedings of the Society of Water Treatment and Examination

20:132123. Brown L, Bellinger EL (1978) Water Res 12:223124. Gorside C (1982) Mar Chem 11:159125. Cox DJ, Hill NC (1985) Estuarine Coastal and Shelf Science 21:127126. Anderson L (1979) Anal Chim Acta 110:123127. Ranger CB (1981) Anal Chem 53:20A128. Ruzicka J, Hansen EH (1978) Anal Chem 99:37129. Johnson KS, Petty RL (1983) Limnol Oceanogr 28:1260130. Hyde DJ, Hill NC (1985) Estuar Coast Shelf Sci 21:127131. Van den Berg CMG (1988) Anal Chim Acta 212:31132. Tyree SY, Bynum MAO (1984) Limnol Oceanogr 29:1337133. Dahiloef I, Svensson O, Tortensson C (1997) J Chromatogr A 771:163134. Sigman DM, Casciotti KL, Andreani C, Barford C, Galanta M, Böhlke JK (2001) Anal

Chem 73:4145135. Sigman DM, Altabet MA, Michener RH, McCorkle DC, Fry B, Holmes RM (1997) Mar

Chem 57:227136. Amberger A, Schmidt H-L (1987) Geochim Cosmochim Acta 51:2699137. Revesz K, Böhlke JK, Yoshinari Y (1997) Anal Chem 69:4375138. Silva SR, Kendall C, Wilkison DH, Ziegler AC, Chang CCY, Avanzino RJ (2000) J Hydrol

228:22139. Liu KK, Su MJ, Hseuh CR, Gong GC (1996) Mar Chem 54:273140. Zumft WG (1997) Microbiol Mol Biol Rev 61:553141. Greenberg EP, Becker GE (1997) Can J Microbiol 23:903142. Cerda A, Oms MT, Forteza R, Cerda V (1998) Anal Chim Acta 371:63143. Sakamoto GM, Friederich GE, Service SK, Chavez FP (1996) Deep-Sea Res 43:1763144. Spencer DW, Brewer PG (1970) Crit Rev Solid State Sci 1:409145. Pett SKW (1979) Water Res Technical Report TR 120146. Matsunaga K, Oyama T, Nishimura M (1972) Anal Chim Acta 58:228147. Gianguzza A, Orecchio S (1982) Inquinamento 24:94148. Wada E, Hattori A (1971) Anal Chim Acta 56:233149. Fogg AG, Bsebsu NK, Abdalla MA (1982) Analyst 107:1040150. Poeston T, Zainal K, Anderson S, Bury SJ (1998) Rapid Commun Mass Spec 12:423151. Van den Berg CMG, Li H (1988) Anal Chim Acta 212:31152. Aminot A, Kerouel R (1998) Mar Chem 61:203153. Bajic SJ, Jaselskis B (1985) Talanta 32:115154. Nagashima K, Matsumoto M, Suzuki S (1985) Anal Chem 57:2065155. Departmentof theEnvironment/National WaterCouncil StudyCommitteeofAnalysts,

Methods for the examination of water and associated materials, HMSO, London (1981)61 pp (RP2B:CENV)

156. Mullin JB, Riley JP (1955) Anal Chim Acta 12:464157. Chow TJ, Johnstone MS (1962) Anal Chim Acta 27:441158. Bernard M, Macchi G, Automation in analytical chemistry, Technicon Symposia, Me-

diad, New York (1965) 252159. Morris AW, Riley JP (1963) Anal Chim Acta 29:272160. Grasshof K (1964) Kieler Meersforsch 20:5161. Wood ED, Armstrong FA, Richards FA (1967) J Mar Biol Assoc 47:23162. Henricksen A, Selmer-Olsen AR (1970) Analyst 95:514


163. Armstrong EAJ, Sterus CR, Strickland JDH (1967) Deep-Sea Res 14:381164. Chow TJ, Johnstone MS (1962) Anal Chim Acta 27:441165. Morris AW, Riley JP (1963) Anal Chim Acta 29:272166. Ruzicka J, Hansen EH (1975) Anal Chim Acta 78:145167. Nydahl F (1976) Talanta 23:349168. Anderson L (1979) Anal Chim Acta 110:123169. Shinn MB (1941) Ind Eng Chem Anal Ed 13:33170. Hansen EH, Ruzicka J (1976) Anal Chim Acta 87:353171. Strickland JDH, Parsons TR (1968) A Practical Handbook of Seawater Analysis, Fish-

eries Research Board of Canada Bulletin 167172. Timmer A, Hoor T (1974) Mar Chem 2:149173. David ARJ, MacCormack T, Morris AW, Worsfold PJ (1998) Anal Chim Acta 31:63174. Methods for theExaminationofWaterandAssociatedMaterials –MethodDContinuous

Flow Methods for the Determination of Total Oxidised Nitrogen or Nitrite in Seawater,Department of the Environment, National Water Council Standing Committee ofAnalysts, HMSO, London (1981)

175. Ito K, Ariyoshi Y, Tanabiki F, Sunbara H (1991) Anal Chem 63:273176. Matthews AD, Riley JP (1970) Anal Chim Acta 51:455177. Johnson KS, Petty RL (1982) Anal Chem 54:1185178. Murphy J, Riley JP (1962) Anal Chim Acta 27:31179. Pakalus P, McAllister BR (1972) J Mar Res 30:305180. Hosokawa I, Ohshima F (1973) Water Res 7:283181. Isaeva AB (1969) Zh Anal Khim 24:1854182. Ducret L, Drouillas M (1959) Anal Chim Acta 21:86183. Sudakov FP, Klitina VI, Dan’shova T Ya (1966) Zh Anal Khim 21:1333184. Babko AK, Shkaravskii Yu F, Kulik VI (1966) Zh Anal Khim 21:196185. Matsuo T, Shida J, Kurihara W (1977) Anal Chim Acta 91:385186. Kirkbright GS, Narayanaswamy R, West TS (1971) Anal Chem 43:1434187. Motomizu S, Wakimoto T, Toei K (1982) Anal Chem 138:329188. Motomizu S, Wakimoto T, Toei K (1984) Talanta 31:235189. Bajic SJ, Jaselkis B (1985) Talanta 32:115190. Hydes DK, Hill NC (1985) Estuar Coast Shelf Sci 21:127191. Flynn WW, Meehan W (1973) Anal Chim Acta 63:483192. Deans DR (1982) Anal Chem 43:2026193. Yoshimura K, Ishii M, Tarutani T (1986) Anal Chem 58:591194. Eberlein K, Kattner G (1987) Fresen Z Anal Chem 326:354195. Airey D, Dal Pont G, Sanders (1984) Anal Chim Acta 166:79196. Itoh K, Nakayama M, Chikuma M, Tanaka H (1985) Fresen Z Anal Chem 321:56197. Wollast R in The Sea, Volume 4, Goldberg E (ed) Wiley New York (1974) p. 359198. Edmund JM, Jacobs SS, Gorden AL (1979) J Geophys Res 84:7809199. Hargis LG (1970) Anal Chim Acta 52:1200. Halasz A, Polak K, Pungor E (1971) Talanta 18:691201. Golkowsa A, Pszonicki L (1973) Talanta 20:749202. Kreshkov AP, Organesyan LB, Ogareva MB (1976) J Anal Chem (USSR) 31:723203. Bashikdash’yan MT, Vasil’eva MG (1981) J Anal Chem (USSR) 36:747204. Riley JP (1975) Chemical Oceanography Volume 3, 2nd Edition Riley JP, Skirrow G (ed)

Academic Press London p. 193205. Armstrong FA (1951) J Mar Biol Assoc 30:149206. Grasshof K (1966) Z Anal Chem 220:89207. Brewer PJ, Riley JP (1966) Anal Chim Acta 35:514

References 97

208. Spencer DW, Brewer PG (1970) Crit Rev Solid State Sci 1:410209. Liss PS, Spencer CP (1969) J Mar Biol Assoc 49:589210. Strickland P, Parsons O (1965) Bulletin of the Fisheries Research Board, Canada 125:50211. Armstrong AJ (1951) J Mar Biol Assoc 30:149212. Grasshof O (1964) Oceanographic Abstracts 11:597213. Thomson J, Johnson KS, Petty RL (1983) Anal Chem 55:2378214. Brzelinski MA, Nelson DM (1986) Mar Chem 19:139215. Hioki A, Law JWH, McLaren JW (1997) Anal Chem 69:21216. Spencer CP in Chemical Oceanography Volume 2, 2nd Edition (1975) Riley JP, Skirow

G (eds) Academic Press New York p250-251217. Sakai H, Fujiwara T, Kumamaru T (1993) Bull Chem Soc Japan 66:3401218. Kanno S, Goto R (1994) Kogyo Yosui 433:50219. Sakai H, Fujiwara T, Kumamaru T (1995) Anal Chim Acta 302:173220. Cutter GA, Oatts TJ (1987) Anal Chem 59:717221. Leck C, Bagander LE (1988) Anal Chem 60:1680222. Fukushi K, Hiiro K (1987) J Chrom 393:433223. Jagner D (1970) Anal Chim Acta 52:483224. Lebel J (1980) Mar Chem 9:237225. Miles DL, Cook JM (1982) Anal Chim Acta 141:207226. Luther GW, Mavarson AL (1975) Anal Chem 47:2058227. Singh RP, Pambid ER, Abbas NM (1991) Anal Chem 63:1897228. Nasu T, Kan M(1988) Analyst 113:1683229. Johnson DL, Pilson EQ (1972) Anal Chem 58:289230. Airey D, Dal Pont G, Sanders G (1984) Anal Chim Acta 166:79231. Hu W, Takeuchi T, Haraguchi H (1993) Anal Chem 65:2204232. Hu W, Haraguchi H (1994) Anal Chem 66:765233. Hu W, Tominaga M, Tao H, Itoh A, Umemura T, Haraguchi H (1995) 67:3713234. Hu W, Haraguchi H (1996) J Chrom 723:251235. Hu W, Hasebe K, Reynolds DM et al. (1997) J Liq Chrom Rel Tech 20:1903236. Hu W, Haddad PR (1998) Trends Anal Chem 17:69237. Hu W, Hasebe K, Reynolds DM, Haraguchi H (1997) Anal Chim Acta 358:143238. Hu W, Haddad PR (1998) Anal Comm 35:317239. Hu W, Haddad PR, Hasebe K, Tanaka K, Tong P, Khoo C (1999) Anal Chem 71:1617240. Sigman DM, Altabet MA, Michener R et al. (1997) Mar Chem 57:227241. Debaar HJW, Van Leeuwe MA, Scharek R, Goeyens L, Bakker KMJ, Fritsche P (1997)

Deep-Sea Res Part II 44:229242. Hutchins DA, Bruland KW (1998) Nature 393:561243. Kirkwood DS, Aminot A, Perttila M (1991) Fourth Intercomparision Exercise for Nu-

trients in Sea Water (ICES) Copenhagen Cooperative Research Report p. 174244. Aminot A, Kerouel R (1991) Anal Chim Acta 248:277245. Aminot A, Kerouel R (1995) Mar Chem 49:221246. Sedijl M, Lu GN (1998) Measuring Science 9:1587247. Yaakov SB, Ruth E (1974) Limnol Oceanogr 19:144248. Yaakov SB, Ruth E (1969) Water Pollution Abstract 42:1757249. Khoo KH, Ramette W, Culbeeson CH, Bates RG (1977) Anal Chem 49:29250. Willie SN, Sturgeon RE, Berman SS (1983) Anal Chem 55:981251. Miyazaki A, Brancho K (1987) Anal Chim Acta 198:297252. Shriadah HMA, Katoeka M, Ohzebi K (1985) Analyst 110:125253. Kuroda R, Matsumoto M, Ogmura K (1988) Fresen Z Anal Chem 330:111

3 Anions in Estuary and Coastal Waters

Very limitedworkhasbeenpublishedon thedeterminationof anions in estuaryand coastal waters. However, there is no doubt that the open seawater analysismethods discussed in Chap. 2 would in many instances be applicable to estuaryand coastal waters.

3.1Nitrate

3.1.1Ultraviolet Spectroscopy

Brown and Bellinger [1] have proposed an ultraviolet technique which is appli-cable to both polluted and unpolluted fresh waters and some estuarine waters.Humic acid and other organics are removed on an ion exchange resin. Bro-mide interference in seawater samples can be minimised by suitable dilutionof the sample but this raises the lower limit of detection such that only on rela-tively rich (0.5 mg/l nitrate) estuarine and inshore waters could the method beapplied. Chloride at concentrations in excess of 10 000 mg/l does not interfere.

The method is either not affected by or can allow for interference from phos-phate, sulfate, carbonate, bicarbonate, nitrite, coloured metal complexes, am-monia, dyes, detergents, phenols, and other ultraviolet absorbing substances.

The determination of nitrate is also discussed under multianion analysis inSect. 3.6.1.

3.2Nitrate and Nitrite

3.2.1Autoanalyser Method

Petts [2] has described a procedure for the determination of total nitrateplus nitrite in estuary waters ranging in salinity from 2.17 to 33.1 g/kg. Inthis method oxidised nitrogen in the sample is reduced to nitrite by a cop-

100 3 Anions in Estuary and Coastal Waters

per/cadmium reduction column and reacted with N-1-naphthylethylene di-aminedihydrochloride under acidic conditions to produce an azo dye whichis evaluated spectrophotometrically at 550 nm. The range of application of themethod is 1–800 µg nitrogen. Standard deviations range from 4.2 µg N/l at the14 µg N/l level to 2.8 µg N/l at the 70 µg N/l level.

3.3Phosphate


A commonly used procedure for the determination of phosphate in seawaterand estuarine waters involves the formation of the molybdenum blue complexat 35–40 ◦Cinanautoanalyser, andspectrophotometric evaluationof thecolourproduced [3]. Unfortunately, when applied to seawater samples, depending onthe chloride content of the sample, peak distortion or even negative peaksoccur which make it impossible to obtain reliable phosphate values. This effectcan be overcome by the replacement of the distilled water used in such methodsby a solution of sodium chloride of appropriate concentration related to thechloride concentration of the sample. The chloride content of the wash solutionneednotbe exactly equal to thatof the sample. For chloride contents ina sampleup to 18 000 mg/l, (i.e., seawater), the chloride concentration in the wash shouldbewithin±15%of that in the sample.Theuseof saline standards is optional butthe use of saline control solutions is mandatory. Using good equipment, downto 0.02 mg/l phosphate can be determined by such procedures. For chloridecontents above 18 000 mg/l, the chloride content of the wash should be within±5% of that in the sample. See also Sect. 3.6.1.

3.4Selenate and Selenite

3.4.1Spectrofluorometric Method

Itoh et al. [4] determined selenium(IV) in estuarine waters by an anion ex-change resin modified with bismuthiol(II) and diaminonaphthalene. An Am-berlite IRA-400 anion exchange resin was modified by mixing with an aqueoussolution of bismuthiol(II) to give 0.2 mmol bismuthiol per g of resin. Sam-ple solution was eluted through a column packed with the modified resin.Selenium(IV) adsorbed as selenotrisulfate was then eluted from the columnwith either 0.1 mol/l penicillamine or 0.1 mol/l cysteine. The eluate was thensubjected to an acid digestion procedure to reduce selenium to the tetravalent

3.5 Sulfate 101

state with diaminonaphthalene for fluorometric determination. Approximateagreement with the tellurium coprecipitation method was obtained. The ap-plication of both methods to the analysis of estuarine waters permitted theseparate determination of both selenium(IV) and selenium(VI), since the tel-lurium coprecipitation method did not differentiate between the two species.


Most cations and anions cause interference on the atomisation of selenium.Bismuth and antimony depress selenium adsorption.

Arsenic lowered somewhat the peak temperature in atomisation profile forselenium. Copper tends to suppress the interferences of diverse elements onatomisation of selenium. However, the interferences from large concentrationsof diverse elements and matrices were not improved even in the presence ofcopper at the atomisation step. Therefore, the separation of selenium frommatrices was recommended.

Selenium is extracted as diethyldithiocarbamate complex from the solutioncontaining citrate and EDTA [5]. Ohta and Suzuki [6] found that only a few el-ements, such as copper, bismuth, arsenic, antimony, and tellurium, are also ex-tracted together with selenium. They examined this for effects of hundredfoldamounts of elements co-extracted with the selenium diethyldithiocarbamatecomplex. An appreciable improvement of interferences from diverse elementswas observed in the presence of copper. Silver depressed the selenium absorp-tion in the case of atomisation of diethyldithiocarbamate complex, but theinterference of silver was suppressed in the presence of copper. The atomisa-tion profile from diethyldithiocarbamate complex was identical with that fromselenide.

No selenium(VI) is extractable from citrate buffered solution containingEDTA and copper. Therefore, tetra- and hexavalent selenium could be deter-mined separately.

The recovery of selenium was satisfactory. The forms of selenium in watersare known to be selenite and selenate [7]. Selenium occurs in non-saline waterat concentrations ranging from less than 0.0002 µg/l to greater than 50 µg/l.Therefore, a large sample size is necessary for analysis at lower concentrationlevels.

3.5Sulfate


The determination of sulfate is discussed under multianion analysis in Sect.3.6.1.

102 3 Anions in Estuary and Coastal Waters

3.6Multianion Analysis

3.6.1Spectrophotometric Method, Sulfate, Phosphate, Nitrate, and Sulfide

Airey et al. [8] have described a method for removing sulfide prior to thedetermination of these anions in anoxic estuarine waters. Mercury(II) chlo-ride was used to precipitate free sulfide from samples of anoxic water. Thesulfide-free supernatant solution was used to estimate sulfide by measuringthe concentration of unreacted mercury(II), as well as to determine sulfate,inorganic phosphate, and nitrate by spectrophotometric methods, in whichsulfide interferes. Sulfide concentrations in the range 0.5–180 000 µg/l sulfurcould be measured, while the lower limits for sulfate, ammonia, nitrite, andinorganic phosphate were 0.024, –1.0 and 1 µg/l, respectively.

References

1. Brown L, Bellinger EL (1978) Water Res 223, 122. Petts KW (1979) The determination of oxidised nitrogen and nitrite in estuarine wa-

ters by autoanalysis, Technical Report TR 120. Water Research Centre, MedmenhamLaboratory, UK

3. Crompton TR, Unpublished work4. Itoh K, Nakayama M, Chikuma M, Tanaka H (1985) Fresen Z Anal Chem 321:565. Bode H (1954) Fresen Z Anal Chem 143:1826. Ohta K, Suzuki M (1980) Fresen Z Anal Chem 302:1777. Cheam V, Agemian H (1980) Anal Chim Acta 113:2378. Airey D, Dal Pont G, Sanders G (1984) Anal Chim Acta 106:79

4 Dissolved Gases

Spencer and Brewer [1] have reviewed methods for the determination of dis-solved gases in seawater (O2, CO2, N2, H2S, inert gases).

4.1Free Chlorine

4.1.1Amperometric Titration Procedures

Although the amperometric titrimetric method has been accepted as a stan-dard method for the determination of total residual chlorine in chlorinatedeffluents [2], recent reports [3,4] have suggested that in the case of chlorinatedwaters, significant errors may occur if certain precautions are not taken. Fur-thermore, somewhat opposing views were presented in these reports on whatthe optimal procedure might be.

When molecular chlorine or hypochlorous acid is added to seawater, thefollowing rapid reactions (reactions (4.1) and (4.2)) ensue and hypobromite isformed [5]:

Cl2 + H2O� H+ + Cl– + HClO (4.1)

HClO + Br– → HBrO + Cl– (4.2)

There is enough bromide in seawater at 35‰ salinity to convert 60 mg/l ofchlorine to hypobromite.

In theamperometric titration for thedeterminationof total residual chlorinein seawater, tri-iodide ions are generated by the reaction between hypochloriteand/or hypobromite with excess iodide pH 4 (reactions (4.3) and (4.4)). The pHis buffered by adding a pH 4 acetic acid–sodium acetate buffer to the sample.

HXO + H+ + 2I– → I2 + H2O + X– (4.3)

I2 + I– → I–3 (4.4)

where X may be Cl or Br. The tri-iodide ion is then titrated with phenylarsineoxide (reaction (4.5)) and the end-point is determined amperometrically.

C6H5AsO + I–3 + 2H2O → C6H5As(OH)2O + 2H+ + 3I– (4.5)

104 4 Dissolved Gases

Carpenter et al. [3] suggest that this method may have underestimated thetrue value by a factor of three or more. The error was proposed to be due tothe oxidation of iodide to iodate by molecular bromine and/or hypobromousacid (reaction (4.6)),

3HBrO + I– → IO–3 + 3H+ + 3Br– (4.6)

and iodate does not revert directly with phenylarsine oxide. Furthermore, atpH 4, the reaction between iodate and excess iodide to form tri-iodide ion(reaction (4.7)) is sluggish:

IO–3 + 8I– + 6H+ → 3I–

3 + 3H2O (4.7)

Thus an apparent loss in total residual chlorine may be observed.Carpenter et al. [3] proposed that, in order to overcome this difficulty,

higher acidity and higher potassium iodide concentrations or a back titrationprocedure should be used.

Goldman et al. [4], on the other hand, suggest that the order of the additionof the reagents for generating tri-iodide ions is crucial for obtaining accurateresults. If the acidic buffer is added first, at pH 4, molecular bromine may beformed (reaction (4.8)):

HBrO + 2H+ + Br– → Br2 + H2O (4.8)

and its loss by volatilisation may cause an apparent loss in total residualchlorine. Thus they recommended that potassium iodide should be addedbefore the acidic buffer solution. When potassium iodide is added to thechlorinated solution, hypoiodite will be formed (reaction (4.9)).

HXO + I– → HIO + X (4.9)

The disproportionation of hypoiodite to form iodate (reaction (4.10)) isbelieved to be slow in slightly alkaline solutions such as seawater [6, 7].

3IO– → IO–3 + 2I– (4.10)

Upon subsequent acidification at pH 4, hypoiodite is converted to moleculariodine (reaction (4.11))

2HIO + 2H+ → 2H2O + I2 (4.11)

which is, in turn, converted to tri-iodide ion (reaction (4.4)). Since tri-iodideis a heavy ion, its loss by volatilization will be negligible if the titration followsthe addition of the reagents promptly. The accepted standard procedure [2]calls for the addition of potassium iodide before the acidic buffer solution.However, this point is not stressed, and its importance is never discussed.Goldman et al. [4] further reported that the backward titration procedure doesnot significantly increase the total residual chlorine concentration.

Since reaction (4.6) is favoured in basic solutions, if Carpenter et al. [3] arecorrect, the longer molecular bromine and/or hypboromous acid are allowed

4.1 Free Chlorine 105

to react with iodide at the natural pH 8 of seawater, the greater the amount ofiodate that will be formed. Consequently, there will be a greater apparent lossof total residual chlorine. This would imply that adding the acidic buffer priorto the addition of potassium iodide may minimise the source of error, becausein an acidic solution reaction (4.6) will be forced to the left. On the other handif acid is added first, the apparent loss of total residual chlorine should increasewith faster stirring rate and with longer time for stirring between the additionof the two reagents. The apparently contradictory implications for these twohypotheses were tested by Wong [8].

Wong [8] found that the determination of residual chlorine in seawaterby the amperometic titrimetic method, potassium iodide must be added tothe sample before the addition of the pH 4 buffer, and the addition of thesetwo reagents should not be more than a minute apart. Serious analyticalerror may arise if the order of addition of the reagents is reversed. Thereis no evidence suggesting the formation of iodate by the reaction betweenhypobromite and iodide. Concentrations of residual chlorine below 1 mg/liodate, which occurs naturally in seawater, causes serious analytical uncer-tainties.

In the sodium borate solution containing bromide, when the pH 4 buffer isadded before the potassium iodate solution, titrations give low total residualchlorine concentrations. This loss increases with the amount of stirring timebetween the addition of the reagents. Even for a stirring time of 10 seconds,there is a loss of about 17% of the total residual chlorine. If the solutionwere stirred for 30 min, 85% of the chlorine would have disappeared. Theconcentration of total residual chlorine determined by the reference methodsdoes not change throughout the experiment. This implies that this loss ofchlorine does not occur in the reaction vessel, but in the titration cell asa result of the analytical procedure.

If potassium iodide is added first, and then the solution is stirred, acidifiedand titrated, the loss of residual chlorine is reduced, although still significant.The loss again increases monotonically with stirring for 20 min. However,for a stirring time of 1 minute or less, the loss is not detectable within theuncertainty of the analytical method. There is a loss of chlorine whether thesample is stirred in the titrator or on a stirrer, although the loss seems smallerin the latter case. For a stirring time of 20 min, only 24% of the residual chlorineis lost. Moreover, titrations performed at pH 2 and pH 4 yield the same residualchlorine concentrations.

Wong [8] reported that the losses of chlorine are not related to the formationof iodate by the oxidation of iodide by hypobromite. The presence of iodatein seawater may cause significant uncertainty in the determination of smallquantities of residual chlorine in water. Determinations of residual chlorine atthe 0.01 mg/l level are of questionable significance.

Carpenter and Smith [9] and Wong [8] pointed out that iodate is alsoa natural constituent of seawater. Thus it may cause a variable positive blank


of up to 0.1 mg/l (1.4 µequiv/l) residual chlorine. This source of error may besafely neglected only for concentrations above 2 mg/l (28 µequiv/l). Carpenterand Smith proposed that iodate may be converted completely to tri-iodide byusing a lower pH or larger excess of iodide in the titration. However, otherside reactions such as the oxidation of iodide by nitrite may interfere with theanalysis under these modified conditions. As this presumably standard methodis being applied to samples with lower and lower concentrations, interferencesand blanks previously considered minor may become significant.

Wong [10,11] has studied this in further detail and found that carrying outthe titration at pH 2 yields a true concentration of total residual chlorine aftercorrection for naturally occurring iodate. The effectiveness of sulfamic acid inthis method for removal of the nitrite interference is shown in Fig. 4.1. In thisexperiment, all the solutions contained 30 µmol/l nitrite, and about 0.5 µmol/lof iodate. The absorbance of the solution at 353 nm decreased with increasingamounts of added sulphamic acid. A constant absorbance was recorded when3 ml or more of 1% (w/v) sulphamic acid was added to the solution, and thisabsorbance was identical with that in a sample containing the same amountof iodate and no nitrite. A concentration of nitrite of 30 µmol/l is unlikely tooccur in estuarine water and seawater:

NH2SO–3 + NO–

2 → N2 + SO2–4 + H2O

Carlson and Weberg [12] have also studied the interference from iodate dur-ing the iodometic determination of residual chlorine in seawater. These work-ers confirmed that due to the presence of naturally occurring iodate, resultsfrom the iodometric determination carried out at pH 2 were up to 20% higher

Figure 4.1. Absorbance of tri-iodide at353 nmformed fromafixedconcentrationof iodate in the presence of 30 µmol/l ni-trite upon the addition of various volumesof 1% (w/v) sulphamic acid; N denotes thecase where neither nitrite nor sulphamicacid was present in the solution. From [8]

4.1 Free Chlorine 107

Figure 4.2. Determination of iodate, bro-mate, and bromide. From [12]

than those obtained in determinations carried out at pH 4. These workers seemto have been unaware of the results obtained by Wong as discussed above.

Carlson and Weberg [12] used mobile phase ion-chromatography to sepa-rate iodate, bromate, bromide, and iodide (Fig. 4.2). Ion-chromatography ofa mixture of iodide and bromide which had been treated with hypochloriteclearly showed the presence of iodate and the absence of bromate. These resultsdemonstrated that iodate generates iodine from iodide very quickly at pH 2(IO–

3 + 5I– + 6H+ → 3I2 + 3H2O) and relatively slowly at pH 4; while bromatefails to reactwith iodideatpH4anddoes soslowlyatpH2(BrO–

3 + 6I– + 6H+ =Br– + 3I2 + 3H2O). Moreover, chromatographic analysis (C18 column) of aniodate solution (10–4 M) after iodide addition (24 min pH 4) confirmed thecontinued presence of iodate. The concentration of iodate after this period(40% of original) agrees with the corresponding titration results.

The practical application of these observations is to minimise the effect ofiodate by rapidly carrying out the iodometric titration of chlorine residualin seawater at pH 4. Moreover, if desired, a titration correction curve can begenerated using iodate at the specific concentration of iodide in the sample inquestion, as there appears to be a complete conversion of seawater iodide toiodate in the presence of excess chlorine.

Other methods for the determination of chlorine in seawater or saline watersare based on the use of barbituric acid [13] and on the use of residual chlorineelectrodes [14] or amperometric membrane probes [15, 16]. In the barbituricacid method [12], chlorine reacts rapidly in the presence of bromide and hascompletely disappeared after 1 minute. This result, which was verified in therange pH 7.5–9.4, proves the absence of free chlorine in seawater. A study ofthe colorimetric deterioration of free halogens by the diethylparaphenylene-diamine technique shows that the titration curve of the compound obtainedis more like the bromine curve than that of chlorine. The author suggests


drawing a calibration curve for different concentrations of chlorine in solutioncontaining 60 mg/l of bromide.

The work carried out by Dimmock and Midgley [14–16] at the CentralElectricity Board, UK is concerned with the analysis of cooling waters, salineand non-saline. Although the analysis of seawater is not specifically discussed,their work may be of some relevance.

4.2Ozone

Sheeter [17] has discussed an ultraviolet method for the measurement ofozone in seawater. Crecelius [18] has discussed oxidation products obtained(bromine, hypobromous acid, bromate) when bromides in seawater are oxi-dised by ozone.

4.3Nitric Oxide

Cohen [19] used electron capture gas chromatography to determine traces ofdissolved nitric oxide in seawater. Precision and accuracy are, respectively, 2%and 3%.

4.4Hydrogen Sulfide

Flow injection analysis has been used for the automated determination of hy-drogen sulfide in seawater [20]. A low-sensitivity flow injection analysis mani-fold for concentrations up to 200 µmol/l hydrogen sulfide had a detection limitof 0.12 µmol/l. Sulfide standards were calibrated by colorimetric measurementof the excess tri-iodide ion remaining after reaction of sulfide with iodine.The coefficient of variation was less than 1% at concentrations greater than10 µmol/l. The method was fast, accurate, sensitive enough for most naturalwaters, and could be used both for discrete and continuous analysis.

4.5Carbon Dioxide

Murphy et al. [21] have described an infrared based detector method for thedetermination of carbon dioxide in seawater.

Neill et al. [22] have described a headspace gas chromatographic methodfor the determination of carbon dioxide (fugacity) in seawater. This methodrequires a small water sample (60 ml), and provides for rapid analysis (2 min).

References 109

Lehaitre et al. [23] have described a fibre-optic spectrophotometric methodfor the in situ measurement of biological and chemical species. This instru-ment can measure phytoplanktonic species, and the potential for chemicalmeasurement such as dissolved carbon dioxide is analysed.

Fukushi and Hiiro et al. [24] described a method for determining totalcarbon dioxide in seawater by capillary isotachoelectrophoresis following iso-lation of the carbon dioxide by membrane permeation.

Tabacco et al. [25] have described a method based on emission spectrometryat an excitation wavelength of 488 nm for the telemetric measurements ofcarbon dioxide tensions in seawater.

Yong Suk Choi et al. [26] used a carbonate-selective electrode to determinedissolved carbon dioxide in the oceans.

References

1. Spencer DW, Brewer PC (1970) Crit Rev Solid State Sci 1:4092. Standard Methods for the Examination of Water and Waste Water Rand MC, Greenberg

AE,TarasMJ(eds)14thEdn(1976)AmericanPublicHealthAssociationWashington,DC3. Carpenter JH, Moore CA, Macalady DL (1977) Environ Sci Technol 11:9924. Goldman JC, Quinby HL, Capuzzo JM (1979) Water Res 13:3155. Wong GTF, Davidson JA (1977)Water Res 11:9716. Li CH, White CF (1943) J Am Chem Soc 65:3357. Truesdale VW (1974) Deep-Sea Res 21:7618. Wong GTF (1980) Water Res 14:519. Carpenter JH, Smith CA in Water Chlorination – Environmental Impact and Health

Effects Jolley RL, Gorchev H, and Hamilton DH (eds) Volume 2 (1978) Ann ArborScience, Ann Arbor, MI, p. 195–208

10. Wong GTF (1982) Environ Sci Technol 16:78511. Wong GTF (1977) Deep-Sea Res 24:11512. Carlson RF, Weberg RT (1977) Chemosphere 12:12513. Figuet JM (1978) Techniques et Sciences Municipales 73:23914. Dimmock NA, Ridgeley D (1982) Talanta 29:55715. Dimmock NA, Ridgeley D (1979) Water Res 13:110116. Dimmock NA, Ridgeley D (1979) Water Res 13:131717. Sheeter H (1973) Water Res 7:72918. Crecelius EA (1979) Journal of the Fisheries Research Board Canada 36:100619. Cohen Y (1977) Anal Chem 49:123820. Sakamoto O, Arnold CM, Johnson VS, Beehler CL (1986) Limnol Oceanogr 31:89421. Murphy PP, Feely RA, Wanninkhof R (1998) Mar Chem 62:10322. Neill C, Johnson KM, Lewis E, Wallace DWR (1997) Limnol Oceanogr 42:177423. Lehaitre M, Le Bris M, Ouisse A, Belinger A (1997) J Mar Environ Eng 4:8524. Fukushi F, Hiro K (1987) Fresen J Anal Chem 328:24725. Tabacco MB, Uttamiat M, Mc Allister M, Walt DR (1999) Anal Chem 71:154 and

a correction in (1999) Anal Chem 71:148326. Choi YS, Lvova L, Shin JH, Oh SH, Lee CS, Kim BH, Cha GS, Nam H (2002) Anal Chem

74:2435

5 Cations in Seawater

5.1Introduction

Determination of trace metals in seawater represents one of the most challeng-ing tasks in chemical analysis because the parts per billion (ppb) or sub-ppblevels of analyte are very susceptible to matrix interference from alkali oralkaline-earth metals and their associated counterions. For instance, the alkalimetals tend to affect the atomisation and the ionisation equilibrium processin atomic spectroscopy, and the associated counterions such as the chlorideions might be preferentially adsorbed onto the electrode surface to give someundesirable electrochemical side reactions in voltammetric analysis. Thus,most current methods for seawater analysis employ some kind of analyte pre-concentration along with matrix rejection techniques. These preconcentrationtechniques include coprecipitation, solvent extraction, column adsorption,electrodeposition, and Donnan dialysis.

Measurement techniques that can be employed for the determination oftrace metals include atomic absorption spectrometry, anodic stripping voltam-metry, differential pulse cathodic stripping voltammetry, inductively coupledplasma atomic emission spectrometry, liquid chromatography of the metalchelates with ultraviolet-visible absorption and, more recently, inductivelycoupled plasma mass spectrometry.

Many of the published methods for the determination of metals in seawa-ter are concerned with the determination of a single element. Single-elementmethods are discussed firstly in Sects. 5.2–5.73. However, much of the pub-lished work is concerned not only with the determination of a single elementbut with the determination of groups of elements (Sect. 5.74). This is particu-larly so in the case of techniques such as graphite furnace atomic absorptionspectrometry, Zeeman background-corrected atomic absorption spectrome-try, and inductively coupled plasma spectrometry. This also applies to othertechniques, such as voltammetry, polarography, neutron activation analysis,X-ray fluroescence spectroscopy, and isotope dilution techniques.

The background concentrations at which metals occur in seawater are ex-tremely low, and much work has been done on preconcentration procedures inattempts to improve detection limits for these metals. Various preconcentra-

112 5 Cations in Seawater

tion techniques, including hydride generation used before atomic absorptionspectrometry, are discussed.

Methods for determining metals in seawater have been published by theStanding Committee of Analysts (i.e., the blue book series, HMSO, London);they are not reproduced in this book, as they are available elsewhere. Thesemethods are based on chelation of the metals with an organic reagent, followedby atomic absorption spectroscopy.

Pelizzetti [496] has reviewed the analysed seawater, including particles,colloids, and trace elements, along with their speciation (66 references).

5.2Actinium

Lin [1] used coprecipitation with lead sulfate to separate 237-actinium fromsea water samples. The 237-actinium was purified by extraction with HDEHP,and determined by alpha spectrometry via Si (Au) surface barrier detection.The method has a sensitivity of 10–3 pCi g–1 of ashed sample.

5.3Aluminium


Aluminiumhasbeendeterminedby spectrophotometricmethodsusingalumi-non [2,3], oxine [4,5], Eriochrome Cyanine R [6] and Chrome Azurol S [7], flu-orometric methods using Pontachrome Blue BlackR [8,9], Lumogallion [9–12]andsalicylaldehydesemicarbazone[13–17], gaschromatographicmethods [18,19], emission spectroscopy [20], and neutron activation analysis [21, 22].Most of these methods necessitate pretreatment steps and special and ex-pensive instruments, require large volumes of sample solution, and are time-consuming. For instance, although the fluorometric method using Lumogal-lion reported by Hydes and Liss [12] is sensitive and rapid, the fluorescencespectrophotometer used is not as popular an instrument as the spectropho-tometer. Dougan and Wilson [23] have also reported the spectrophotomet-ric determination of aluminium (at concentrations of 0.05 and 0.3 mg/l) inwater with pyrocatechol violet, and Henriksen and coworkers [24, 25] haveimproved the method to some extent, but these procedures are not sat-isfactory for the concentrations of aluminium normally found in seawater(about 2 µg/l).

Korenaga et al. [26] have described an extraction procedure for the spec-trophotometric determination of trace amounts of aluminium in seawaterwith pyrocatechol violet. The extraction of ion-associate between the alu-minium/pyrocatechol violet complex and the quaternary ammonium salt,

5.3 Aluminium 113

zephiramine (tetradecyldimethybenzyl ammonium chloride), is carried outwith 100 ml seawater and 10 ml chloroform. The excess of reagent extracted isremoved by back-washing with 0.25 M sodium bromide solution at pH 9.5.

The calibration graph at 590 nm obeyed Beer’s law over the range 0.13–1.34 µg aluminium. The apparent molar absorptivity in chloroform was9.8 ×104 l mol–1 cm–1.

Several ions (e.g., manganese, iron (II), iron (III), cobalt, nickel, copper,zinc, cadmium, lead, and uranyl) react with pyrocatechol violet, and to someextent are extracted together with aluminium. The interferences from theseions and other metal ions generally present in seawater could be eliminatedby extraction with diethyldithiocarbamate as masking agent. With this agentmost of the metal ions except aluminium were extracted into chloroform, andother metal ions did not react in the amounts commonly found in seawater.Levels of aluminium between 6 and 6.3 µg/l were found in Pacific Ocean andJapan Sea samples by this method.

5.3.2Spectrofluorometric Methods

Howard [27] determined dissolved aluminium in seawater by the micelle-enhanced fluorescence of its lumogallion complex. Several surfactants (toenhance fluorescence and minimise interferences), used for the determina-tion of aluminium at very low concentrations (below 0.5 µg/l) in seawaters,were compared. The surfactants tested in preliminary studies were anionic(sodium lauryl sulfate), non-ionic (Triton X-100, Nonidet P42, NOPCO, andTergital XD), and cationic (cetyltrimethylammonium bromide). Based on thedegree of fluorescence enhancement and ease of use, Triton X-100 was se-lected for further study. Sample solutions (25 ml) in polyethylene bottleswere mixed with acetate buffer (pH 4.7, 2 ml) lumogallion solution (0.02%,0.3 ml) and 1,10-phenanthroline (1.0 ml to mask interferences from iron).Samples were heated to 80 ◦C for 1.5 h, cooled, and shaken with neat sur-factant (0.15 ml) before fluorescence measurements were made. This proce-dure had a detection limit at the 0.02 µg/l level. The method was indepen-dent of salinity and could therefore be used for both freshwater and seawatersamples.

Salgado Ordonez et al. [28] used di-2-pyridylketone 2-furoyl-hydrazone asa reagent for the fluorometric determination of down to 0.2 µg aluminium inseawater. A buffer solution at pH 6.3, and 1 ml of the reagent solution wereadded to the samples containing between 0.25 to 2.50 µg aluminium. Fluores-cence was measured at 465 nm, and the aluminium in the sample determinedeither from a calibration graph prepared under the same conditions or a stan-dard addition procedure. Aluminium could be determined in the 10–100 µg/lrange. The method was satisfactorily applied to spiked and natural seawatersamples.



Spencer and Sachs [29] determined particulate aluminium in seawater byatomic absorption spectrometry. The suspended matter was collected fromseawater (at least 2 litres) on a 0.45 µm membrane filter, the filter was ashed,and the residue was heated to fumes with 2 ml concentrated hydrofluoric acidand one drop of concentrated sulfuric acid. This residue was dissolved in2 ml 2 M hydrochloric acid and the solution was diluted to give an aluminiumconcentration in the range 5–50 µg/l. Atomic absorption determination wascarried out with a nitrous oxide acetylene flame. The effects of calcium, iron,sodium, and sulfate alone and in combination on the aluminium absorptionwere studied.

5.3.4Anodic Stripping Voltammetry

Van den Berg et al. [30] determined aluminium in seawater by anodic strip-ping voltammetry. They give details of a procedure for the determination ofdissolved aluminium in natural waters, including seawater, by complexationwith 1,2-dihydroxyanthraquinone-3-sulfonic acid, collection of the complexon a hanging mercury drop electrode, and determination by cathodic strip-ping voltammetry. The advantages of this method over other techniques areindicated and optimal conditions are described. The total time required was10–15 min per sample and the limit of detection was 1 nmol/l aluminiumfor an adsorption time of 45 s. No serious interferences were found, but ul-traviolet irradiation was recommended for samples with high organic con-tent.


An example of a gas chromatographic method is that of Lee and Burrell [18]. Inthis method the aluminium is extracted by shaking a 30 ml sample (previouslysubjected to UV radiation to destroy organic matter) with 0.1 M trifluoroacety-lacetone in toluene for 1 h. Free reagent is removed from the separated toluenephase by washing it with 0.01 M aqueous ammonia. The toluene phase is in-jected directly onto a glass column (15 cm×6 mm) packed with 4.6% of DC710and 0.2% Carbowax 20 M on Gas-Chrom Z. The column is operated at 118 ◦Cwith nitrogen as carrier gas (285 ml/min) and electron capture detection. Ex-cellent results were obtained from 2 µl of extract containing 6 pg of aluminium.See also Sects. 5.74.9 and 5.74.17.

5.4 Ammonium 115

5.4Ammonium


Various workers have discussed indophenol methods for the determination ofammonium in seawater [31–46].

Haywood and Huyser’s [43] study of the indophenol blue reaction for am-monia determination showed the effect of pH variation on the final colourand emphasised the necessity for the efficient buffering to obtain reproducibleresults. The optimum conditions found by the authors (pH 10.8 and usingsodium nitroprusside) were similar to those used by Solorzano [37]. Haywoodand Huyser also replaced acetone with sodium nitroprusside.

Berg and Addullah [47] have described a spectrophotometric autoanalysermethod based on phenol, sodium hypochlorite, and sodium nitroprusside forthe determination of ammonia in sea and estuarine water (i.e., the indophenolblue method).

The manifold design allows for the determination of ammonia concen-tration in the range 0.2–20 µg/l as NH4 over a salinity range 35–10‰, withnegligible interference from amino acids.

The interference from amino acids was investigated and found to be negligi-ble as reportedbySolorzano [37] andHaywoodandHuyser [43],whoemployedno heating for the indophenol blue colour development. Solutions containing50 µg N/l of urea, histidine, lycine, glycine, and alanine were analysed. TheNH4-N detected ranged between 0.4% (for urea) and 2.2% (for alanine) of thenitrogen added.

Hampson [56] used ferrocyanide catalyst and ultraviolet photon activa-tion energy of appropriate frequency to activate selectively the reaction be-tween ammonia, phenol, and sodium hypochlorite. The reaction is carriedout at an optimum pH of 10.5, since urea may break down at a pH over11 and at a low temperature of 30 ± 1 ◦C. This avoids alkaline hydrolysis ofamino acids to ammonia, a process known to occur extensively at highertemperatures. All conditions, including photon flux and ionic activities, areprecisely controlled to give stability when indophenol blue type methodssuch as those of Koroleff and Solarzano [36, 37] have been applied to cer-tain types of samples. Hampson [56] investigated the causes of such coloursuppression (Fig. 5.1). Very strong suppression is noted, even at low amineconcentrations, and even with indophenol blue formation from the reagentblank suppressed at amine concentrations above 25 ppm N. Numerical anal-ysis of the type of relationship between amine concentration and indophenolblue response to ammonia suggests a third- or possibly fourth-power poly-nomial function (as distinct from a simple power or exponential relation-ship).


Figure5.1. Suppression of indophe-nol blue response to ammoniaby dimethylamine in seawater;0.4 ppm ammonia N in seawater.From [56]

Noneof theseaminesgaveany indophenolblue responsewhenpure seawater(ammonia free) to which they had been added was subjected to the standardanalytical procedure.

Toovercome the suppressioneffectof amines in thedeterminationof ammo-nia, Hampson [56] investigated the effect of nitrite ions added either as nitriteor as nitrous acid. Figure 5.2 indicates that very considerable suppression bynitrite does occur, although it is not as strong as with any of the amines. Again,it is not great so long as the nitrite N concentration is less than the ammonia Nconcentration, but rapidly increases as the nitrite concentration exceeds theammonia concentration. In fact, the nitrite modified method was found to besatisfactory in open seawater samples and polluted estuary waters.

The determination of ammonia in non-saline waters does not present anyanalytical problems and, as seen above, reliable methods are now available forthe determination of ammonia in seawaters. In the case of estuarine waters,however,newproblemspresent themselves.This isbecause thechloridecontentof such waters can vary over a wide range from almost nil in rivers entering theestuary to about 18 g/l in the edges of the estuary where the water is virtuallypure seawater.

Particularly in autoanalyser methods this wide variation in chloride contentof the sample can lead to serious “salt errors” and, indeed, in the extreme case,can lead to negative peaks in samples that are known to contain ammonia.Salt errors originate because of the changes of pH, ionic strength and opticalproperties with salinity. This phenomenon is not limited to ammonia deter-mination by autoanalyser methods; it has, as will be discussed later, also beenobserved in the automated determination of phosphate in estuarine samplesby molybdenum blue methods.

In a typical survey carried out in an estuary, the analyst may be presentedwith several hundred samples with a wide range of chloride contents. Beforestarting any analysis, it is good practice to obtain the electrical conductivitydata for such samples so that they can be grouped into increasing ranges ofconductivity and each group analysed under the most appropriate conditions.

5.4 Ammonium 117

Figure5.2. Suppression of indophe-nol blue response to ammonia bynitrite ion in seawater; 0.4 ppm am-monia N in seawater. Source: [56]

Brezinski [48] has described a spectrophotometric method for the determi-nation of nanomolar concentrations of ammonium in seawater. To seawatersamples (180 ml) was added a sequence of reagent solutions in deionised dis-tilled water as follows: phenol (2.4 ml, 10%), sodium aquopentacyanoferrate(1 ml of a freshly prepared solution containing 0.03 g sodium aquopentacyano-ferrate in 104 ml of double distilled water) and sodium hypochlorite (6 ml,5.5%). The sodium aquopentacyanoferrate acted as a coupling reagent in theformation of indophenol. Reaction mixtures were kept in the dark for 2 h at40 ◦C and allowed to cool for 1.5 hours before adding phosphoric acid (1 M;1.65 ml) and n-hexane (6.6 ml). The organic phase containing indophenol waspipetted into a clean tube and methylene chloride (10 ml) added followedby pH 12 buffer (10 ml). Indophenol blue was re-extracted into the aqueousphase and its concentration determined colorimetrically at 640 nm. Interfer-ence effects by metals, nitrite, urea, and amino acids presents in seawater arediscussed. Calibration curves were linear to 2 µM ammonium.

Le Corre and Treguer [49] developed an automated procedure based onoxidation of the ammonium ion by hypochlorite in the presence of sodiumbromide followed by spectrophotometric determination of the nitrite. Thestandard deviation on a set of samples containing 1µg NH–

4-N per litre was 0.02.This method was compared with an automated method for the determinationof ammonia as indophenol blue. The results from the two methods are in goodagreement.

Urea and amino acids interfere in this procedure. Le Corre and Treguer [49]discuss the effect of salinity on the determination of ammonia and describea suitable correction procedure.

Other workers who have investigated automated methods for the determi-nation of ammonia in seawater include Grasshoff and Johnnsen [50], Berg andAbdullah [47], Truesdale [51], Matsumaga and Nishimura [46].

Selmer and Sorenson [52] have described a procedure for extraction of am-monium from seawater for 15N determinations. In this method ammoniumnitrogen was converted to indophenol and concentrated onto an octadecylsi-


lane column. Subsequent analysis of the indophenol was that for whole cellmaterial and the atom% of 15N was determined by emission spectrometry. Themethod was accurate and precise when coupled with other reported methods.The application of the method to field experiments on the west coast of Swedenis described.


Willason and Johnson [53] have described a modified flow-injection analysisprocedure for ammonia in seawater. Ammonium ions in the sample wereconverted to ammonia which diffused across a hydrophobic membrane andreacted with an acid-based indicator. Change in light transmittance of theacceptor steam produced by the ammonia was measured by a light emittingdiode photometer. The automated method had a detection limit of 0.05 µmol/land a sampling rate of 60 or more measurement per hour.

5.4.3Ion-Selective Electrodes

McLean et al. [54] have applied polarography to the determination of ammoniaand other nitrogen compounds in brine samples, and Gilbert and Clay [55]have investigated the determination of ammonia in seawater using the ammo-nia electrode. These latter workers showed that down to 0.01 ppm ammoniacan be determined using an electode (Orion model 95–10) incorporating a hy-drophobic membrane that separates the sample solution (adjusted to pH 11with sodium hydroxide) from an internal solution 0.1 M in ammonium chlo-ride. A glass pH electrode and a Ag–AgCl reference electrode are immersed inthe aqueous ammonium chloride. The ammonia in the sample passes throughthe membrane and the change in pH in the internal solution is detected bythe glass electrode. The behaviour and characteristics of the system, includingtheoretical limits, are discussed.

5.4.4High-Performance Liquid Chromatography

Gardner et al. [57, 58] have applied this technique to the determination ofammonium ion in seawater. The liquid chromatography method involvedfluorometric detection, after post column labelling with o-phthalaldehyde/2-mercaptoethanol reagent. This method was developed to directly quantify15NH4[14NH4 + 15NH4] ion ratios in aqueous samples that had been enrichedwith 15NH4 for isotope dilution experiments. Cation exchange chromatogra-phy, with a sodium borate buffer mobile phase, was selected as the separation

5.5 Antimony 119

mode because the two isotopes have slightly different constants in the equi-librium reaction between ammonium ion and ammonia. When the two formsof ammonium were passed separately through a high-performance cation ex-change column under precisely controlled chromatographic conditions, theretention time (RT) of 15NH4 was 1.012 times the RT of 14NH4. The two iso-topic forms of ammonium ion were not resolved into separate peaks when theywere injected together, but the retention time of the combined peak, as definedby an integrator, increased as increasing percentages of 15NH4 in the samplewere converted to ammonia, which diffused across a hydrophobic membraneand reacted with an acid-based indicator.

5.5Antimony


Sturgeon et al. [59] have described a hydride generation atomic absorptionspectrometry method for the determination of antimony in seawater. Themethod uses formation of stibene using sodium borohydride. Stibine gas wastrapped on the surface of a pyrolytic graphite coated tube at 250 ◦C and anti-mony determined by atomic absorption spectrometry. An absolute detectionlimit of 0.2 ng was obtained and a concentration detection limit of 0.04 µg/lobtained for 5 ml sample volumes.

5.5.2Hydride Generation Atomic Absorption Spectrometry

Bertine and Lee [60] have described hydride generation techniques for deter-mining total antimony, Sb (V), Sb (III), Sb–S species and organo-antimonyspecies in frozen seawater samples.

Total Antimony

Total antimony content was analysed by a hydride generation technique utilis-ing a quartz burner with a hydrogen flame. The water sample was made 2 N inhydrochloric acid with a final volume of 100 ml. Two ml of 20% (w/v) potassiumiodide were added and the sample degassed using helium as a carrier gas for100 s. A silanised glass wool trap on which to collect the SbH, was then placedin liquid nitrogen and 2 ml of 5% sodium borohydride were slowly injectedover a time period of 100 s. The sample was stripped for 300 s, then the trap wasremoved from the liquid nitrogen and the hydride was carried to an electri-cally heated quartz burner with a hydrogen flame. The antimony concentration


was measured using atomic absorption spectrophotometry. Detection limitsof about 0.01 ng are obtainable.

Both the hydrochloric acid and sodium borohydride contributed to theblank. The Sb (V) in the 12 N hydrochloric acid was removed by uptake ona Dowex 1–X8 anion exchange resin. Sodium borohydride was purified, afterdissolution, by addition of 0.5 ml sodium hydroxide (50%) to 200 ml of 5%sodium borohydride, and subsequent filtration through a hydrochloric acidprecleaned 0.45 µm Millipore membrane.

Hydrogen sulfide gas interfered in the determination of antimony since,after the addition of hydrogen sulfide, a peak comes a few seconds after theantimony peak. It was found that either degassing the sample for 300 s orplacing lead acetate in the line eliminated the problem without interferingwith the antimony determination.

Antimony (III)

Two ml of citrate buffer (purified by an Fe-APDC precipitation) were added tomaintain a pH of 5–6. The sample was degassed for 100 s prior to the injectionof 2 ml of 5% sodium borohydride. The sample was stripped for 300 s. Thisprocedure gives a complete extraction of antimony (III) and no extraction ofantimony (V), even in the presence of hundredfold higher concentrations ofantimony (V).

At a pH of 5–6, hydrogen sulfide evolution by degassing proceeds at a muchslower rate. Background correction using a hydrogen lamp or lead acetateplaced in line was able to remove any interference from the amount remainingafter 400 s. However, it was found that the extraction yield of antimony (III)standards prepared in sodium sulfide was not only incomplete but the yield wasalso inversely proportional to the amount of sulfide added to the standards. Ithas been hypothesised that in sulfide-rich waters an antimony sulfide complexmay exist. A complete yield of antimony (III) in sodium sulfide could beattained by making a 1–2 ml sample 2 N in hydrochloric acid, degassing for5 min, bringing the volume to 100 ml, adding sufficient Tris–buffer to bringthe pH to 6, then proceeding with the hydride generation method as above. Noantimony (V), even in thousandfold excess, was detected by the above method.

See also Sects. 5.74.3, 5.74.7, 5.74.9, 5.74.11, 5.74.14 and 5.74.15.

5.6Arsenic


Afansev et al. [61] have described an extraction photometric method for thedetermination of arsenic at the µg/l range in seawater. This method usesdiantipyrilmethane as the chromogene reagent. The coefficient of variation is

5.6 Arsenic 121

2.5% for arsenic concentrations in the 1.5–5 µg/l range. Good agreement wasobtained with results obtained by neutron activation analysis.

A UK standard official method [62] has been published for the spectropho-tometric determination of arsenic in sea water. The determination is effectedby conversion to arsine using sodium borohydride which is added slowly tothe acidified samples by a peristaltic pump. The liberated arsine is trappedin an iodine/potassium iodide solution and the resultant arsenate determinedspectrophotometically as the arsenomolybdenum blue complex at 866 nm.The method is applicable down to 0.19 µg arsenic.


Bermejo–Barrera et al. [64] studied the use of lanthanum chloride and mag-nesium nitrate as modifiers for the electrothermal atomic spectrometric de-termination of µg/l levels of arsenic in seawater.

Howard and Comber [63] converted arsenic in seawater to its hydride priorto determination by atomic absorption spectrometry.

Electrothermal atomic absorption spectrometry is used to study the totalarsenic and arsenic (III) content in marine sediments [64].

Burton and coworkers [65] have studied the distribution of arsenic in theAtlantic Ocean. Samples from 1000 m and above were filtered through acid-washed 0.45 µm Sartorius membrane filters. Analyses on samples from depthsbelow 1000 m were made on unfiltered water.

Aliquots of 50 ml were placed in a round-bottomed flask, fitted with a mod-ified Dreschel head and an injection syringe in a side arm. Concentratedhydrochloric acid 20 ml, 1 ml of 1 M ascorbic acid solution and 1 ml 1 M potas-sium iodide solution were added. The solution was stood for 30 min to allowreduction of AsV to AsIII which was necessary to ensure quantitative recoveryof inorganic arsenic as arsine under the conditions used in the subsequentstep. With nitrogen passing through the flask at a flow rate of 150 ml/min,0.5 ml 8% w/v sodium borohydride solution was added from the syringe. Thearsine evolved was trapped in 2 ml of a solution containing 0.7% w/v potassiumiodide, over a period of 3 min.

The concentrates were subsequently analysed for arsenic using Varian-TechtronAASatomicabsorptionspectrophotometerfittedwithaPerkin-ElmerHGA 72 carbon furnace, linked to a zinc reductor column for the generationof arsine (Fig. 5.3). A continuous stream of argon was allowed to flow withthe column connected into the inert gas line between the HGA 72 control unitand the inlet to the furnace. Calcium sulfate (10–20 mesh) was used as anadsorbent to prevent water vapour entering the carbon furnace. The carbontube was of 10 mm id and had a single centrally located inlet hole.

A wide range of elements was tested for interfering effects; the only sig-nificant interferences found were at concentrations much higher than those


Figure 5.3. Zinc reactor columnfor generation of arsine by elec-trothermal atomic absorptionspectrophotometry. Source: Au-thor’s own files

encounted in seawater. No significant difference in the results were found whena sample of seawater was analysed in the way described and also by the sameprocedure but using the method of standard additions.

5.6.3Neutron Activation Analysis

The neutron activation method for the determination of arsenic and antimonyin seawater has been described by Ryabin et al. [66]. After coprecipitation ofarsenic acid and antimony in a 100 ml sample of water by adding a solutionof ferric iron (10 mg iron per litre) followed by aqueous ammonia to givea pH of 8.4, the precipitate is filtered off and, together with the filter paper, iswrapped in a polyethylene and aluminium foil. It is then irradiated in a silicaampoule in a neutron flux of 1.8 ×1013 neutrons cm–2 s–1 for 1–2 h. Two daysafter irradiation, the γ -ray activity at 0.56 MeV is measured with use of a NaI(Tl) spectrometer coupled with a multichannel pulse-height analyser, andcompared with that of standards.

5.6 Arsenic 123

Yusov et al. [67] separated arsenic (III) and arsenic (V) in seawater usinga chloroform solution of ammonium pyrrolidine diethyldthiocarbamate. Theseparated fractions were then analysed by neutron activation analysis.

5.6.4Inductively Coupled Plasma Mass Spectrometry

Creed et al. [68] described a hydride generation inductively coupled plasmamass spectrometric method featuring a tubular membrane gas–liquid separa-tor for the determination of down to 100 pg of arsenic in seawater.

Klane and Blum [69] showed that inductively coupled plasma spectrometrywasable todeterminebelow1000 ng/lof arsenic inseawater. Ionexclusionchro-matography coupled with inductively coupled plasma mass spectrometry hasbeen used to determine several arsenic species in seawater [947]. Down to 3 ng/larsenic can be determined using hydride generation prior to this technique.


Jaya et al. [70] carried out an anodic stripping voltammetric determination ofarsenic (III) at a copper-coated glassy electrode. The deposition of copper onthe electrode made it sensitive to the presence of arsenic (III) and suitable foruse by anodic stripping voltammetry analysis. The height of the stripping peakwas linearly dependent on the concentration of arsenic (III) in the solution for7.5–750 µg/l arsenic (III). Lead, zinc, cadmium, manganese, and thallium didnot cause significant interference, but bismuth did. The method gave 92–106%arsenic recovery when tested on synthetic seawater samples, and on naturalarsenic-free seawater spiked with arsenic at levels of 10 and 20 ng per litre.

Hua et al. [71] carried out automated determination of total arsenic inseawater by flow constant-current stripping with gold fibre electrodes in whichthe sample was acidified and pentavalent arsenic was reduced to the trivalentformwith iodide.Thearsenicwas thendepositedpotentiostatically for4minona 25 µm gold fibre electrode, and subsequently stripped with constant currentin 5 M hydrochloric acid. Cleaning and regeneration of the gold electrode werefully automated.

Huiliang et al. [72] have described a flow potentiometric and constant-current stripping analysis for arsenic (V) without prior chemical reductionto arsenic (III). Details are given of procedure for determination of pentava-lent arsenic by means of flow potentiometry and constant-current strippinganalysis. It involved reduction of arsenic to the element state on a gold-platedplatinum fibre electrode at very low reduction potential, and subsequent re-oxidation either by means of a constant current, or chemically using gold asoxidant. Methods for applying this technique to determination of total arsenicin acidified seawater are presented.



Becker et al. [73] have described a method for the determination of dissolvedarsenic in seawater at µg/l levels by precipitation and energy dispersive X-rayfluorescence spectroscopy.Arsenicwasprecipitatedasmagnesiumammoniumarsenatewithmagnesiumammoniumphosphate as carrier.Theprecipitatewascollected on a glass fibre filter. An energy-dispersive X-ray spectrometer witha rhodium primary target operated at 60 kV and 2 mA and a silver secondarytarget was used to measure arsenic. Reagents were optimised for 200 ml sam-ples, and arsenic recovery was greater when 3 ml of phosphate carrier wasused. The limit of detection was 0.7 µg/l. The method was suitable for all typesof natural waters, including seawater.

See also Sects. 5.74.3, 5.74.6, 5.74.7, 5.74.9, 5.74.14, and 5.74.15.

5.7Barium

Barium is of oceanographic interest since it is a nonconservative stable traceelement. In spite of the short 10 000 year oceanic residence time for barium,ocean biology largely determines its distributions in the ocean interior. Dis-solved concentrations in the major oceans-mapped as part of the GEOSECSprogramme lie in the range 20–200 nmol/kg (5.6–28 µg/l), and profiles showthe lowest concentrations near the surface and enrichment at depth in a fashionsimilar (but not identical) to the distribution of the nutrient element silicon.Its determination to a precision of better than 1% by isotope dilution massspectrometry has earned barium the distinction of being the “best measured”nutrient-like trace metal in seawater [74].


Bishop [75] determined barium in seawater by direct injection Zeeman-mod-ulated graphite furnace atomic absorption spectrometry. The V2O3/Si modi-fier added to undiluted seawater samples promotes injection, sample drying,graphite tube life, and the elimination of most seawater components in a slowchar at 1150–1200 ◦C. Atomisation is at 2600 ◦C. Detection is at 553.6 nmand calibration is by peak area. Sensitivity is 0.8 absorbance s/ng (M0 = 5.6 pg0.0044 absorbance s) at an internal argon flow of 60 ml/min. The detection limitis 2.5 pg barium in a 25 ml sample or 0.5 pg using a 135 ml sample. Precisionis 1.2% and accuracy is 23% for natural seawater (5.6–28 µg/l). The methodworks well in organic-rich seawater matrices and sediment porewaters.

Epstein and Zander [76] used graphite furnace atomic absorption spec-trometry for the direct determination of barium in seawater and estuarine

5.8 Beryllium 125

water. Roe and Froelich [81] achieved a detection of 30 pg barium for 50 µl in-jections of seawater using direct injection graphite furnace atomic absorptionspectrometry.

Dehairs et al. [78]describeamethod for the routinedeterminationofbariumin seawater using graphite furnace atomic absorption spectrometry. Bariumis separated from major cations by collection on a cation exchange column.The barium is removed from this resin with nitric acid. Recoveries are greaterthan 99%.

5.8Beryllium

5.8.1Graphite Furnace Atomic Absorption Spectrometry

Okutani et al. [79] have achieved a rapid and simple preconcentration ofberyllium by selective adsorption using activated carbon as an adsorbentand acetylacetone as a complexing agent. The method has been used for thedetermination of a trace amount of beryllium by graphite furnace atomic ab-sorption spectrometry. The beryllium-acetylacetonate complex is adsorbedeasily onto activated carbon at pH 8–10. The activated carbon which adsorbedthe beryllium-acetylacetonate complex was separated and dispersed in purewater. The resulting suspension was introduced directly into the graphite fur-nace atomiser. The determination limit was 0.6 ng/l (S/N = 3), and the relativestandard deviation at 0.25 µg/l was 3.0–4.0% (n = 6). Not only was there nointerference from the major ions such as sodium, potassium, magnesium, cal-cium, chloride, and sulfate in seawater, but there was also no interference fromother minor ions. The method was applied to the determination of nanogramsper millilitre levels of beryllium in seawater and rainwater.

5.8.2Miscellaneous

Other methods reported for the determination of beryllium include UV-visible spectrophotometry [80, 81, 83], gas chromatography (GC) [82], flameatomic absorption spectrometry (AAS) [84–88] and graphite furnace (GF)AAS [89–96]. The ligand acetylacetone (acac) reacts with beryllium to forma beryllium–acac complex, and has been extensively used as an extractingreagent of beryllium. Indeed, the solvent extraction of beryllium as the acety-lacetonate complex in the presence of EDTA has been used as a pretreat-ment method prior to atomic absorption spectrometry [85–87]. Less than1 µg of beryllium can be separated from milligram levels of iron, aluminium,chromium, zinc, copper, manganese, silver, selenium, and uranium by thismethod. See also Sect. 5.74.9.


5.9Bismuth


Shijo et al. [95] converted bismuth in seawater into its dithiocarbamate com-plex, and then extracted the complex into xylene prior to determination inamounts down to 0.3 ppt by electrothermal atomic absorption spectrometry.

Soo [96] determined picogram amounts of bismuth in seawater by flamelessatomic absorption spectrometry with hydride generation. The bismuth is re-duced in solution by sodium borohydride to bismuthine, stripped with heliumgas, and collected in situ in a modified carbon rod atomiser. The collectedbismuth is subsequently atomised by increasing the atomiser temperature anddetected by an atomic absorption spectrophotometer. The absolute detectionlimit is 3 pg of bismuth. The precision of the method is 2.2% for 150 pg and6.7% for 25 pg of bismuth. Concentrations of bismuth found in the PacificOcean ranged from < 0.003–0.085 (dissolved) and 0.13–0.2 ng/l (total).

Gilbert and Hume [97], Florence [98], and Eskilsson and Jaguar [99] haveapplied anodic stripping voltammetry to the determination of bismuth inseawater. Gilbert and Hume [97] and Florence [98] investigated the electro-analytic chemistry of bismuth (III) in the marine environment using linear-sweep anodic stripping voltammetry and a film of mercury on a glassy car-bon [98] or a graphite [97] substrate as working electrode. Gillain et al. [100]used differential-pulse anodic stripping voltammetry with a hanging mer-cury drop electrode for the simultaneous determination of antimony (III) andbismuth (III) in seawater.

In the method of Gilbert and Hume [97], the sample contained in a sil-ica cell was purged and stirred by passage of purified nitrogen. A platinumcounter-electrode was used. The reference electrode consisted of a silver wire,previously anodised in seawater, held in a borosilicate glass tube containinga small untreated portion of the sample that was separated from the samplebeing analysed by a plug of unfused Vycor. To diminish the effect of the steeplyrising background current (0.1 µA s–1) on the stripping peaks, a compensatingcircuit was devised. Bismuth was deposited at –0.4 V from seawater made 1 Min hydrochloric acid and gave a stripping peak of –0.2 V, the height of whichwas proportional to concentration without interference from antimony or met-als normally present. Antimony was deposited at –0.5 V from seawater made4 M in hydrochloric acid and gave a stripping peak at –0.3 V, the area of whichwas proportional to the sum of antimony and bismuth. By use of the standard-addition technique, satisfactory results were obtained for the concentrationranges 0.2–0.09 µg kg–1 for bismuth and 0.2–0.5 µg kg–1 for antimony.

Florence [98] carried out anodic stripping voltammetry of bismuth ina weakly acidic medium, with a polished vitreous carbon electrode mercury-

5.10 Boron 127

plated in situ. The limit of detection is 5 ng bismuth per litre. Seawater wasfound to contain 0.02–0.11 µg bismuth per litre in surface samples.

Since computerised potentiometric stripping analysis [101–103] is in manyrespects a simpler analytical technique than linear-sweep or differential-pulseanodic stripping voltammetry, it is used for the potentiometric stripping de-termination of bismuth (III). Although only data for the determination ofbismuth (III) in seawater are reported in this paper, the optimum experimen-tal conditions with respect to sample matrix, interferences, limits of detection,and other experimental parameters can be applied to samples other than salinewaters. During the course of his work it became apparent that the surface Kat-tegatt samples analysed during this investigation contained approximately oneorder of magnitude less bismuth (III) than the results obtained hitherto by elec-troanalytical [97,98,100] and ion exchange [104] techniques. Because the directdetermination of such low concentrations of bismuth in seawater by means ofpotentiometric stripping analysis would be somewhat time-consuming, a sim-ple preconcentration technique was used. This technique was based on thecoprecipitation of bismuth (III) with magnesium hydroxide, thus taking ad-vantage of the naturally high magnesium concentration in seawater [105,106].

Square-wave anodic stripping voltammetry was employed by Komorsky-Lovric [107] for the determination of bismuth in seawater. A bare glassy-carbon rotating disk electrode was preconditioned at –0.8 V versus Ag/AgCl,prior to concentration of bismuth. The method was applied to seawater in the12 ng/l range.

See also Sects. 5.74.5, 5.74.7, 5.74.8, 5.74.10, and 5.74.11.

5.10Boron


Various chromogenic reagents have been used for the spectrophotometricdetermination of boron in seawater. These include curcumin [108, 109], nileblue [110], and more recently 3,5 di-tert butylcatechol and ethyl violet [111].Uppstroem [108] added anhydrous acetic acid (1 ml) and propionic anhydride(3 ml) to the aqueous sample (0.5 ml) containing up to 5 mg of boron per litreas H3BO3 in a polyethylene beaker. After mixing and the dropwise additionof oxalyl chloride (0.25 ml) to catalyse the removal of water, the mixture isset aside for 15–30 minutes and cooled to room temperature. Subsequently,concentrated sulfuric–anhydrous acetic acid (1:1) (3 ml) and curcumin reagent(125 mg curcumin in 100 ml anhydrous acetic acid) (3 ml) are added, and themixed solution is set aside for at least 30 minutes. Finally 20 ml standard buffersolution (90 ml of 96% ethanol, 180 g ammonium acetate – to destroy excessof protonated curcumin – and 135 ml anhydrous acetic acid diluted to 1 litre


with water) is added, the mixture is cooled to room temperature, and theextinction is measured at 545 nm. For less than 0.01 mg boron per litre, thecoloured complex must be concentrated: a portion of sample (2–10 ml) inwhich the colour reaction has taken place is diluted with water (100 ml), andthe complex is extracted into 5 or 10 ml of extractant (100 ml isobutyl methylketone, 150 ml of chloroform, and 1 g phenol). The extinction of the organicphase is measured at 545 nm. The colour of the complex is stable for about 2 h.Interference is caused by germanium and fluoride. Small amounts of water aretolerated, but they reduced the efficiency of the method.

A curcumin-based automated version of the above procedure [78] has beendescribed [79]. Determinations can be made in the range 0.1–6 mg boronper litre. At a level of 3 mg/l the coefficient of variation was 1.5% and thedetection limit was 0.01 mg/l. Up to 240 samples per hour can be processed bythe procedure.

In the Nile blue spectrophotometric method, 10 ml 2%aqueous hydrofluoricacid is added to a 10 ml sample contained in a polyethylene bottle. The mixtureis shaken for about 2 h. Aqueous ferrous sulfate 10% 10 ml and 1 ml 0.1%aqueous Nile blue A are added, then extracted with o-dichlorobenzene (10 mland 3 × 5 ml). The combined organic extracts are diluted to 50 ml with thesolvent and the extinction measured at 647 nm. Interference from chloride ionsup to 100 mg/l can be eliminated by precipitation as silver chloride.

Marcantoncetos et al. [112] have described a phosphorimetric method forthe determination of traces of boron in seawater. This method is based on theobservation that in the “glass” formed by ethyl ether containing 8% of sulfuricacid at 77 K, boric acid gives luminescent complexes with dibenzoylmethane.A 0.5 ml sample is diluted with 10 ml 96% sulfuric acid, and to 0.05–0.3 ml ofthis solution 0.1 ml 0.04 M dibenzoylmethane in 96% sulfuric acid is added.The solution is diluted to 0.4 ml with 96% sulfuric acid, heated at 70 ◦C for 1 h,cooled, ethyl ether added in small portions to give a total volume of 5 ml, andthe emission measured at 77 K at 508 nm, with excitation at 402 nm. At the levelof 22 ng boron per ml, hundredfold excesses of 33 ionic species give errors ofless than 10%. However, tungsten and molybdenum both interfere.


Atomic absorption spectrometry has been used for the rapid determination ofboron in seawater [113].

5.10.3Coulometry

Tsaikov [114] has described a coulometric method for the determination ofboron in coastal seawaters. This method is based on the potentiometric titra-

5.11 Cadmium 129

tion of boron with electrogenerated hydroxyl ions, after removal of the cationcomponents by ion exchange. The method has good reproducibility and ismore accurate than other methods; it is also fairly rapid (25–30 minutes perdetermination).

5.11Cadmium

In the determination of cadmium in seawater, for both operational reasons andease of interpretation of the results it is necessary to separate particulate mate-rial from the sample immediately after collection. The “dissolved” trace metalremaining will usually exist in a variety of states of complexation and possiblyalso of oxidation. These may respond differently in the method, except wheredirect analysis is possible with a technique using high-energy excitation, suchthat there is no discrimination between different states of the metal. The onlytechnique of this type with sufficiently low detection limits is carbon furnaceatomic absorption spectrometry, which is subject to interference effects fromthe large and varying content of dissolved salts.


Various workers have discussed the application of graphite furnace atomicabsorption spectrometry to the determination of cadmium in seawater [115–124].

Batley and Farrah [120] and Gardner and Yates [118] used ozone to decom-pose organic matter in samples and thus break down metal complexes prior toatomic absorption spectrometry. By this treatment, metal complexes of humicacid and EDTA were broken down in less than 2 min. These observations ledGardner and Yates [118] to propose the following method for the determinationof cadmium in seawater.

The sample is filtered immediately after collection, acidified to about pH 2,and transferred to a 1-litre Pyrex storage bottle. Prior to extraction the sampleis ozonised in the sample bottle for 30 minutes. Nitrogen is passed through thesample for 5 minutes to remove excess ozone, then the pH is carefully raisedto about 5 by addition of ammonia solution, and about 5 ml Chelex 100 resinin the ammonia form is added. After stripping for at least 1 h, the resin iscollected in a Pyrex chromatography column and washed with the calculatedquantityofanappropriatebuffer toelute calciumandmagnesium.After furtherwashing with 50 ml deionised water, the resin is eluted with 2 M nitric acid toa volume of 25 ml. The elute is analysed by graphite furnace atomic absorptionspectrometry.

Danielson et al. [119] have described a method for the determination ofcadmium in seawater. The samples were analysed by graphite furnace atomic


absorption spectroscopy after a two-stage extraction. Replacing the acetatebuffer and performing the extraction in a clean room with Teflon utensilssignificantly improved blank levels. Extractions were performed on boardship immediately after sampling, and the extracts brought home for analysis.An aliquot of the sample was also transferred into carefully cleaned TeflonFEP bottles and acidified with 1 ml of nitric acid per litre. The nitric acid waspurified by sub-boiling distillation. These samples were extracted about twomonths after sampling at the shore laboratory. The same method was usedwith the exception that extra ammonia was added to the buffer to compensatefor the acidification. The method was applied to arctic seawaters, and showeda profile of cadmium with sampling depth ranging from 0.133 nmol/l cadmiumat the surface to 0.205 nmol/l cadmium at 2000 m.

As cadmium is one of the most sensitive graphite furnace atomic absorptiondeterminations, it is not surprising that this is the method of choice for thedetermination of cadmium in seawater. Earlier workers separated cadmiumfrom the seawater salt matrix prior to analysis. Chelation and extraction [121–128], ion exchange [113, 124, 125, 129], and electrodeposition [130, 131] haveall been studied.

The direct determination of cadmium in seawater is particularly diffi-cult because the alkali and alkaline earth salts cannot be fully charred awayat temperatures that will not also volatilise cadmium. Most workers in thepast [125, 132–135] who have attempted a direct method have volatilised thecadmium at temperatures which would leave sea salts in the furnace. Thisrequired careful setting of temperatures, and was disturbed by situations thatcaused temperature settings to change with the life of the furnace tubes.

Lundgren et al. [132] showed that the cadmium signal could be separatedfrom a 2% sodium chloride signal by atomising at 820 ◦C, below the tem-perature at which the sodium chloride was vaporised. This technique hasbeen called selective volatilisation. They detected 0.03 µg/l cadmium in the 2%sodium chloride solution. They used an infrared optical temperature monitorto set the atomisation temperature accurately.

Campbell and Ottaway [136] also used selective volatilisation of the cad-mium analyte to determine cadmium in seawater. They could detect 0.04 µg/lcadmium (2 pg in 50 µl) in seawater. They dried at 100 ◦C and atomised at1500 ◦C with no char step. Cadmium was lost above 350 ◦C. They could notuse ammonium nitrate because the char temperature required to remove theammonium nitrate also volatilised the cadmium. Sodium nitrate and sodiumand magnesium chloride salts provided reduced signals for cadmium at muchlower concentrations than their concentration in seawater if the atomisationtemperature was in excess of 1800 ◦C. The determination required lower atom-isation temperatures to avoid atomising the salts. Even this left the magnesiuminterference, which required the method of additions.

Guevremont et al. [117] used a direct, selective volatilisation determinationof cadmium in seawater. They used 20 µl seawater samples, 1 g/l of EDTA, an

5.11 Cadmium 131

atomisation ramp from 250 ◦C to 2500 ◦C in 5 s, and the method of additions.Their detection limit was 0.01 µg/l (0.2 pg in 20 µl), and the characteristicamount was 0.7 pg/0.0044 A. The EDTA promoted the early atomisation ofcadmium below 600 ◦C. Their test seawater sample (0.053 µg/l) was confirmedby other methods. These authors were able to separate reliably the cadmiumand background signals by using the method of Campbell and Ottaway [136];the EDTA made this possible.

Guevremont et al. [117] studied the use of various matrix modifiers inthe graphite furnace gas method of determination of cadmium in seawater.These included citric acid, lactic acid, aspartic acid, histidine, and EDTA. Theadditionof less than1 mgofanyof the compounds to1 ml seawater significantlydecreased matrix interference. Citric acid achieved the highest sensitivity andreduction of interference, with a detection limit of 0.01 µg cadmium per litre.

In similar work, Sturgeon et al. [125] compared direct furnace methodswith extraction methods for cadmium in coastal seawater samples. They couldmeasure cadmium down to 0.1 µg/l. They used 10 µg/l ascorbic acid as a matrixmodifier. Various organic matrix modifiers were studied by Guevremont [116]for this analysis. He found citric acid to be somewhat preferable to EDTA,aspartic acid, lactic acid, and histidine. The method of standard additionswas required. The standard deviation was better than 0.01 µg/l in a seawatersample containing 0.07 µg/l. Generally, he charred at 300 ◦C and atomised at1500 ◦C. This method required compromise between char and atomisationtemperatures, sensitivity, heating rates, and so on, but the analytical resultsseemed precise and accurate. Nitrate added as sodium nitrate delayed thecadmium peak and suppressed the cadmium signal.

Sperling [133] has reported extensively on the determination of cadmiumin seawater, as well as in other biological samples and materials. He addedammonium persulfate, which permitted charring seawater at 430 ◦C withoutloss of cadmium. For work below 2 µg/l cadmium in seawater he recommendedextraction of the cadmium to separate it from the matrix [126, 134, 135]. Hefound no change in the measured levels over many months when the seawaterwas stored in high-density polyethylene or polypropylene.

Pruszkowska et al. [135] described a simple and direct method for the de-termination of cadmium in coastal water utilizing a platform graphite furnaceand Zeeman background correction. The furnace conditions are summarisedin Table 5.1. These workers obtained a detection limit of 0.013 µg/l in 12 µlsamples, or about 0.16 pg cadmium in the coastal seawater sample. The char-acteristic integrated amount was 0.35 pg cadmium per 0.0044 A s. A matrixmodifier containing di-ammonium hydrogen phosphate and nitric acid wasused. Concentrations of cadmium in coastal seawater were calculated directlyfrom a calibration curve. Standards contained sodium chloride and the samematrix modifier as the samples. No interference from the matrix was observed.

Seawater samples usually contain a total of 2–3% of several alkali and alka-line earth salts, with sodium chloride as a main constituent. A 2 µl sample of


Table 5.1. Zeeman graphite Furnace Conditions

Dry Char Atomise Clean-out Cool

Temperature, ◦C 160 550 1600 2600 20Ramp, s 1 1 0 1 1Hold, s 60 45 5 6 20Internal gas flow, ml/min 300 300 0 300 300Recorder, s –5

From [135]

seawater charred at 700 ◦C has a background signal so high that even theZeeman correction system cannot handle it (Fig. 5.4). The largest amountsof sodium choride present in seawater are reportedly volatilised below 950 ◦C[137], but even with ammonium phosphate, the matrix modifier recommendedfor cadmium, it is notpossible to char at sohigha temperature. Figure 5.4 showsthat 200 µg of di-ammonium hydrogen phosphate reduced the SB signal of

Figure 5.4. Background (SB) profiles for 2 µl seawater alone, with 200 µg (NH4)2HPO4 andwith 500 µg NH4NO3. The char temperature was 700 ◦C and the atomisation temperaturewas 1700 ◦C. The signals from the Data System 10 reported here are called ZAA signals forthe analytical result, and SB (single beam) signals for the backgrounds. The SB signals areexpressed in absorbance units (A) and the ZAA signals are usually in absorbance unit –seconds (A-s). The SB signal is signal plus background, but for the small analyte signals ofthis study, the SB signal is effectively background. The actual integrated absorbance signalsthat were used were calculated by software on the Data Station 10 from signals transmittedby the Zeeman/5000. The plots in later figures show typical signals, but were not used forquantitative evaluation. Source: [135]

5.11 Cadmium 133

2 µl seawater to 0.5 A, but 500 µg ammonium nitrate reduced the backgroundmore effectively to 0.16 A. No reduction of the cadmium signal occurred inthe presence of ammonium nitrate if the char temperature was below 600 ◦C,and phosphate was used as a matrix modifier. If ammonium nitrate was usedwithout phosphate, the cadmium was lost at temperatures below 500 ◦C. Theaddition of the phosphate stabilised the cadmium, while the ammonium nitratepromoted the release or conversion of the bulk of the material responsible forthe background. The addition of the phosphate produced a background signalthat appeared much later than the cadmium peak.

It was shown that 1.25 mg ammonium nitrate is enough to keep the back-ground signal below 1.5 A, and there are no large differences in backgroundabsorbances for amounts from 1.25 to 7.5 mg ammonium nitrate.

It was also shown that 2% nitric acid reduced the background to a level thatcan be handled by the Zeeman correction system. From 4% to 8% nitric acid,the changes in background signal shapes were not very large.

Typical sample (ZAA) signals and background (SB) signals for a seawatersample are shown in Fig. 5.5. Brewer [138] has used electrically vaporisedthin gold film atomic emission spectrometry to determine cadmium at the10 ppb level in highly acidic saline solutions following preconcentration ona strong-base anion exchange resin.

Knowles [139] used extraction with ammonium pyrollidine dithiocarba-mate dissolved in methyl isobutyl ketone to extract cadmium from seawater

Figure 5.5. Zeeman profiles of a seawater sample (Sandy Cove N.9) and Sb profiles. The firstpair of profiles represents a single 12 µl aliquot, the second pair, two aliquots, and the thirdpair, three aliquots. The modifier was 200 µg (NH4)2HPO4, 8% HNO3, and 5 µg Mg(NO3)2.The char temperature was 550 ◦C, and the atomisation temperature 1600 ◦C. Source: [135]


prior to analysis by Zeeman atomic absorption spectrometry. The method wascapable of determining 0.04 µg/l of cadmium in seawater when concentrationfactors of 100 were used.

Three Zeeman-based methods for the determination of cadmium in sea-water were investigated. Direct determinations can be made with or withoutthe use of a pyrolytic platform atomisation technique. The wall atomisationmethods presented were considerably faster than the platform atomisationtechnique. For extremely low levels of cadmium, indirect methods of analysisemploying a preliminary analyte extraction can be employed. Background lev-els are minimal in extracted samples, and the total furnace programme timewas the lowest of the methods examined.

Lum and Callaghan [140] did not use matrix modification in the electother-mal atomic absorption spectrophotometric determination of cadmium in sea-water. The undiluted seawater was analysed directly with the aid of Zeemaneffect background correction. The limit of detection was 2 ng/l.

Electrothermal atomic absorption spectrophotometry with Zeeman back-ground correction was used by Zhang et al. [141] for the determination ofcadmium in seawater. Citric acid was used as an organic matrix modifier andwas found to be more effective than EDTA or ascorbic acid. The organic ma-trix modifier reduced the interferences from salts and other trace metals andgave a linear calibration curve for cadmium at concentrations ≤ 1.6 µg/l. Themethod has a limit of detection of 0.019 µg/l of cadmium and recoveries of95–105% at the 0.2 µg of cadmium level.

Lum and Callaghan [140] determined down to 2 ng/l of cadmium directlyin seawater by atomic absorption spectrometry with Zeeman correction.

Han et al. [142] have reported an atomic absorption spectrometric methodfor the determination of cadmium in seawater using sodium phosphate formatrix modification.


Stolzberg [143] has reviewed the potential inaccuracies of anodic strippingvoltammetry and differential pulse polarography in determining trace metalspeciation, and thereby bio-availability and transport properties of trace met-als in natural waters. In particular it is stressed that nonuniform distribution ofmetal–ligand species within the polarographic cell represents another limita-tion inherent in electrochemical measurement of speciation. Examples relateto the differential pulse polarographic behaviour of cadmium complexes ofNTA and EDTA in seawater.

In a method described by Yoshimura and Uzawa [144], cadmium in seawateris coprecipitated with zirconium hydroxide (Zr(OH)4) prior to determinationby square-wave polarography. The precipitate is dissolved in hydrochloricacid, and cadmium concentration is determined from the peak height of the

5.12 Caesium 135

polarogram at –0.64 V. The calibration curve was linear for concentrations of≤ 5.0 µg of cadmium.

Kounaves and Zirino [145] studied cadmium–EDTA complex formation inseawater using computer-assisted stripping polarography. They showed thatthe method is capable of determining the chemical speciation of cadmium inseawater at concentrations down to 10–8 M.

Turner et al. [146] studied the automated electrochemical stripping of cad-mium in seawater.

See Sects. 5.74.4–5.74.6, 5.74.8–5.74.12, and 5.74.14–5.74.16.

5.12Caesium

Nuclear activities such as electricity production by nuclear power plants, oraccidents such as occurred at Chernobyl, release radionulides, including cae-sium, into the environment. The caesium concentrations in these matrices isvery low, so that in addition to a sensitive analytical method, it is necessaryto make use of an enrichment technique to bring the caesium concentrationwithin the scope of the analytical method.


Atomic absorption spectrometry is suitable as a method of analysis of theconcentrate, and is applicable to radioactive and non-radioactive forms of theelement.

For the enrichment of caesium from seawater and other types of sample,Frigieri et al. [147] used ammonium hexcyanocobalt ferrate. This was chosenbecause it can be employed in strongly acidic solutions, with the exception ofconcentrated sulfuric acid.

Atomicabsorptionspectrometryhasbeenused todeterminecaesiuminsea-water. The method uses preliminary chromatographic separation on a strongcation exchange resin, ammonium hexcyanocobalt ferrate, followed by elec-trothermal atomic absorption spectrometry. The procedure is convenient, ver-satile, and reliable, although decomposition products from the exchanger,namely iron and cobalt, can cause interference.

Caesium is fully retained by a chromatographic column of ammoniumhexcyanocobalt ferrate, and can be recovered by dissolution of the ammoniumhexcyanocobalt ferrate in hot 12 M sulfuric acid.

As iron and cobalt both interfere with the determination of caesium, usingthe 852.1 nm caesium line, these elements were removed in a preliminaryseparation and then caesium determined.

Ganzerli et al. [148] also used copper hexacyanoferrate (II) on a silica sup-port to absorb caesium from both seawater and fresh water. A specific analyt-


ical method is not described, although atomic absorption spectrophotometrymight be used.

Shen and Li [149] extracted caesium (and rubidium) from brine sampleswith 4-tert-butyl-2-(α methyl-benzyl) phenol prior to atomic absorption spec-trometric determination of the metal.

See also Sect. 5.74.16.

5.13Cerium

See Sects. 5.49 and 5.74.15.

5.14Calcium

5.14.1Titration Methods

Jagner [150] used computerised photometric titeration in a high-precisiondetermination of calcium in seawater.

Calcium is titrated with EGTA (1,2-bis-(2-amminoethoxyethane N,N,N′,N′–tetracetic acid) in the presence of the zinc complex of zincon as indirect indi-cator for calcium. Theoretical titration curves are calculated by means of thecomputer program HALTAFALL in order to assess accuracy and precision. Themethod gives a relative precision of 0.00028 when applied to estuarine waterof 0.05–0.35% salinity.

Complexometric titration is considered to be the best method for the de-termination of calcium, but investigators have differed in the endpoint de-tection technique used and their evaluation of interference by other alkalineearth elements. Studies using different endpoint techniques, some of whichalso considered magnesium to calcium ratios in seawater, do not agree onthe effect of magnesium on the titration of calcium with EGTA (1,2-bis-(2-amminoethoxyethane NNN1N1–tetra-acetic acid). Table 5.2 lists the findingsof some of these studies; the reference cited reports that magnesium eitherhas no effect, causes a positive interference, and, in one case, has a negativeinterference.

In most cases where strontium interference was evaluated, a positive inter-ference was found, but the degree of correction (of the calcium titre) variedfrom about –0.38% in several studies to –0.7% and –0.88% in other investiga-tions which claim that all or nearly all strontium is co-titrated.

In the light of these observations Olson and Chen [164] decided to usea correction factor for use in their visual endpoint calcium titration methodinvolving titration with EGTA. They found that interferences by magnesium

5.14 Calcium 137

Table 5.2. Reported studies on determination of calcium using EDTA

Conclusion Method Reference

No Mg interference at seawater ratios Hg electrode [151]Positive Mg interference fromMg:Ca = 1.5 and higher

Zn-Zincon [152]

No Mg interference at seawater ratios Zn-Zincon [153]Titration error if Mg > Ca Theoretical, Zn-Zincon [154]No Mg interference at seawater ratioswhen end-points sharp

Various chemical visualindicators

[155]

Mg interference of +0.792% on Ca titre; Srinterference of +0.388% on Ca Titre

Zn-Zincon [156]

Mg interference of –0.23% on Ca titre; Srinterference of +0.77% on Ca titre

GHA [157]

Sr interference; increased titration errorat seawater ratios – dependent on end-point sensitivity

Stability constants ‘conditionalconstants’

[158]

Sr interference of +0.37% on Ca titre Ca-Red [159]Mg interference at seawater ratios Computer simulated curves of

Zn-Zincon[160]

No Mg interference; Sr interference of0.9% on Ca titre

Amalgamated Ag electrode [161]

No Mg interference Ca ion selective [162]No Mg interference; No Sr interference Ca ion selective [163]

Source: own files

and strontium were insignificant at the molar ratios normally found in seawa-ter, but is more serious in samples containing higher ratios of magnesium orstrontium to calcium. An average value of 0.02103 was obtained for the ratioof calcium to chlorinity in samples of standard seawater.

They used the titration method of Tsunogai et al. [157]. The titrant solu-tions were standardised against calcium carbonate of primary standard quality(99.9975% purity) rather than zinc, and the EGTA (Eastman Chemicals) wasused without further purification.

The presence of normal concentrations of sodium, magnesium, and stron-tium have no net effect on the determination of calcium above the approximatelevel of accuracy of about 0.1% so that no correction factor seems necessary.A sufficient amount of titrant must be added to complex at least 98% of dis-solved calcium before the buffer is added; this apparently reduces the loss ofcalcium by coprecipitation with magnesium hydroxide.

Interference effects begin to appear at higher magnesium or strontium mo-lar ratios. Tsunogai et al. [157] found the interference of magnesium to benegative and, for strontium, interference is related to the extraction into theorganic layer of the calcium GHA complex. They found a positive interferencefor strontium at twice the seawater molar ratios. Therefore, the interference of


the individual alkaline earth elements on the calcium titration found by Olsonand Chen [164] are consistent in direction, though clearly not in magnitudewith those that were reported Tsungai et al. [157]. The presence of sodium(chloride) in the solutions also seems to diminish these interference effects inboth cases. Although no explanation was found for the reduced interferenceeffect when sodium is present, it does suggest the advantage of either standar-dising the titrant against seawater matrix calcium standard or of having somematrix available to evaluate individual interference effects with a procedure tobe used for seawater.


Atomic absorption spectrophotometry [165, 166] has been used in the deter-mination of calcium and magnesium in seawater.

5.14.3Flame Photometry

Blake et al. [167] have described a flame photometric method for the deter-mination of calcium in solutions of high sodium content. The method wasapplied to simulated seawater. In the method Chelex 100 chelating resin (Na+

form) (20 g) is stirred with 2 N hydrochloric acid (15 ml) for 5 min, the acidis decanted and the resin is washed with water (2 × 25 ml), stirred with 2 Nsodium hydroxide (15 ml) for 5 min and again washed with water (2 × 25 ml).The procedure is repeated five times, then the resin is dried at 100 ◦C. A neu-tral solution (100 ml) containing up to 50 ppm of calcium and up to 4% ofsodium is passed through a column of the resin, a specified amount of hy-drochloric acid (pH 2.4) is passed through, and the percolate containing thesodium is discarded. Elution is then effected with 2 N hydrochloric acid (5 ml)and the column is washed with water (25 ml), the combined eluate and wash-ings are diluted to 100 ml and calcium is determined by flame photometryat 622 nm. There is no interference from magnesium, zinc, nickel, barium,mercury, manganese, copper or iron present separately in concentrations of25 ppm or collectively in concentrations of 5 ppm each. Aluminium depressesthe amount of calcium found.

5.14.4Calcium-Selective Electrodes

Whitfield et al. [168] used a calcium-selective electrode to monitor EGTA andDCTA titrations of aqueous mixtures of calcium, magnesium, and sodium.The concentrations were selected to span the range of natural waters and the

5.15 Chromium 139

results were analysed statistically. The pattern of titration curves observedwith changing solution composition agreed qualitatively with that predictedtheoretically, but the overall potential drop was usually lower than that pre-dicted; endpoints were determined by graphical and numerical methods. Thetechnique is suitable for the determination of calcium and magnesium in sea-water with an estimated accuracy of 0.5%. The electrode also responds to zinc,iron, lead, copper, nickel, and barium. In seawater free from coastal influ-ences the concentration of these elements are too low to cause interference.However, in inshore samples these elements might have to be masked. Thetitration of calcium in other natural waters by this method should give anaccuracy of 1 to 2% which could be improved by adding known amounts ofcalcium to bring the initial concentration in the sample to a suitable value(e.g. 10–2 M).


Brenner et al. [169] applied inductively coupled plasma atomic emission spec-trometry to the determination of calcium (and sulfate) in brines. The principaladvantage of the technique was that it avoided tedious matrix matching ofcalibration standards when sulfate was determined indirectly by flame tech-niques. It also avoided time-consuming sample handling when the sampleswere processed by the gravimetric method. The detection limit was 70 µg/land a linear dynamic range of 1 g/l was obtained for sulfate.

See also Sects. 5.74.1, 5.74.2, 5.74.14 and 5.74.16.

5.15Chromium

5.15.1Total Chromium

Reported concentrations of chromium in open ocean waters range from 0.07to 0.96 µg/l with a preponderance of values near the lower limit. Methodsused for the determination of chromium at this concentration have gener-ally used some form of matrix separation and analyte concentration prior todetermination [170–173], electroreduction [174, 175] and ion exchange tech-niques [176, 177].

Whereas it is desirous to utilise analytical schemes that permit elucidationof the various chromium species particularly since CrVI is acknowledged tobe a toxic form of this element, it is also useful to have the capability of rapid,total chromium measurement.

Determinationof chromiumbymanyof themethodscitedabove isproblem-atic. Variable and nonquantitative recovery with chelation–solvent extraction


techniques necessitates use of the method of additions [173]. Coprecipitationtechniques require lengthy processing times and extensive sample manipula-tion. Ion exchange suffers from slow uptake and release kinetics, necessitatingtotal destruction and solubilisation of the resin [177] or complex apparatusand multicomponent eluting solutions.

Atomic Absorption Spectrometry

As a consequence Willie et al. [178] developed an approach to the determina-tion of total chromium. This involves preliminary concentration of dissolvedchromium from seawater by means of an immobilised diphenylcarbazonechelating agent, prior to determination by atomic absorption spectrome-try. A Perkin-Elmer Model 500 atomic absorption spectrometer fitted witha HGA-500 furnace with Zeeman background correction capability was usedfor chromium determinations. Chromium was first reduced to CrIII by addi-tion of 0.5 ml aqueous sulfur dioxide and allowing the solution to stand forseveral minutes. Aliquots of seawater were then adjusted to pH 9.0 ± 0.2 byusing high-purity ammonium hydroxide and gravity fed through a column ofsilica at a nominal flow rate of 10 ml/min.

The sequestered chromium was then eluted from the column with 10.0 ml0.2 M nitric acid. More than 93% of chromium was recovered in the first 5 ml ofeluate by this method. Extraction of 80 ng spikes of CrIII from 200 ml aliquotsof seawater was quantitative. Neither CrIII nor CrVI could be quantitativelyextracted. Between 0.15 and 0.19 µg/l total chromium was found in seawaterby this method compared to the accepted value of 0.184 ± 0.016 µg/l.

Moffett [179] determined chromium in seawater by Zeeman correctedgraphite tube atomisation atomic absorption spectrometry. The chromium isfirst complexed with a pentan-2,4 dione solution of ammonium 1 pyrrolidinecarbodithioc acid, then this complex extracted from the water with a ketonicsolvent such as methyl isobutyl ketone, 4-methylpentan-2-one or diisobutylketone.

Recoveries obtained on standard samples are adequate, e.g., US EPA stan-dard 4, quoted 10.2 ± 1.1 ng/l, found, 9.75 ± 0.16 ng/l.

Chemiluminescence

Dubovenko et al. [180] used chemiluminescence to determine total chromiumin brines. The method is based on the enhancement of the chemilumines-cence by chromium in the reaction of 4-(diethylamino) phthalhydrazide withhydrogen peroxide. The detection limit is 0.025 µg/l of chromium, and thechemiluminescence is directly proportional to chromium concentrations inthe range 5 ×10–10 to 10–6 M.

5.15 Chromium 141

Isotope Dilution – Mass Spectrometry

Isotopedilutiongaschromatography-mass spectrometryhasalsobeenused forthedeterminationofppbof total chromiumin seawater [181–183].The sampleswere reduced to ensure CrIII and then extracted and concentrated as tris(1,1,1-trifluoro-2,4-pentanediono) chromium (III) [(Cr(tfa)3)] into hexane.The Cr(tfa)+

2 mass fragments were monitored into a selected ion monitoring(SIM) mode.

Isotope dilution techniques are attractive because they do not require quan-titative recovery of the analyte. One must, however, be able to monitor specificisotopes which is possible by using mass spectrometry.

In this method, chromium is extracted and preconcentrated from seawaterwith trifluoroacetylacetone [H(tfa)] which complexes with trivalent but nothexavalent chromium. Chromium reacts with trifluoroacetylacetone in a 1:3ratio to form an octahedral complex, Cr(tfa)3. The isotopic abundance of itsmost abundant mass fragment, Cr(tfa)+

2 was monitored by a quadrupole massspectrometer.

A mass spectrum of Cr(tfa)3 is shown in Fig. 5.6. The isotopic distributionof the Cr(tfa)+

2 fragment (m/e 358 and 359 here) is evident. This is readilycalculable if the individual elemental abundances are known. Assuming theisotopic abundance of 12C and 13C to be 98.89% and 1.11 and 50Cr, 52Cr, 53Crand 54Crbe4.31, 83.76, 9.55and2.38%, respectively, andneglectingany isotopicabundances less than 1%, one can obtain a set of calculated abundances forthe Cr(tfa)+

2 ion. The agreement between the two sets is excellent. Agreement

Figure 5.6. Mass spectrum of Cr(tra)3. Source: Author’s own files


with data obtained by isotope dilution spark source mass spectrometry [184]and graphite furnace [185] was excellent.

5.15.2Chromium (III)

Anodic Stripping Voltammetry

Boussemart and Van den Berg [186] adsorbed chromium (III) in seawater ontosilica, then reoxidised it to chromium (VI) prior to determination in amountsdown to 1 pmol/l by a voltammetric procedure.

The chemiluminescence technique has been used to determine trivalentchromium in seawater. Chang et al. [187] showed Luminol techniques for de-termination of chromium (III) were hampered by a salt interference, mainlydue to magnesium ions. Elimination of this interference is achieved by seawa-ter dilution and utilising bromide ion chemiluminescence signal enhancement(Fig. 5.7). The chemiluminescence results were comparable with those ob-tained by a graphite furnace flameless atomic absorption analysis for the totalchromium present in samples. The detection limit is 3.3 ×10–9 mol/l (0.2 ppb)for seawater with a salinity of 35%, with 0.5 M bromide enhancement.

The effect of calcium interference is somewhat different. At its concentra-tion in seawater, 0.010 M, calcium ion had no effect upon chemiluminescenceanalysis of a 6 ×10–8 M CrIII solution in the absence of bromide ion. The

Figure 5.7. MgII interference of Clanalysis for Cr. -◦-◦-◦ = in thepresence of 0.3 Br–; –•–•–• =in the absence of Br–. CrIII =6 ×10–8 M. EDTA = 2.5 ×10–3 M.Source: [187]

5.15 Chromium 143

chemiluminescence signal dropped to zero, however, if the calcium ion con-centration was increased to 0.013 M. In the presence of 0.3 M bromide ion,no interference was observed for analysis of 6 ×10–8 M CrIII when the calciumconcentration was less than or equal to 0.002 M. The chemiluminescence signalincreased linearly with increasing calcium ion concentration when the calciumconcentration exceeded 0.002 M.

The combined effect of cation interference for both MgII and CaII is almostidentical with the solid curve in Fig. 5.7, indicating that the magnesium ioninterference is the dominant one.

5.15.3Chromium (III) and (VI)

Various workers have discussed the separate determination of CrIII and CrVI

in seawater [170–172, 188, 189].

Spectrophotometric Method

Diphenylcarbazone and diphenylcarbazide have been widely used for the spec-trophotometric determination of chromium [190]. CrIII reacts with diphenyl-carbazone whereas CrVI reacts (probably via a redox reaction combined withcomplexation) with diphenylcarbazide [191]. Although speciation would seema likely prospect with such reactions, commercial diphenylcarbazone is a com-plex mixture of several components, including diphenylcarbazide, diphenyl-carbazone, phenylsemicarbazide, and diphenylcarbadiazone, with no stoi-chiometric relationship between the diphenylcarbazone and diphenylcarba-zide [192]. As a consequence, use of diphenylcarbazone to chelate CrIII selec-tively also results in the sequestration of some CrVI. Total chromium can be de-termined with diphenylcarbazone following reduction of all chromium to CrIII.

Use of immobilised chelating agents for sequestering trace metals fromaqueous and saline media presents several significant advantages over che-lation–solvent extraction approaches to this problem [193, 194]. With littlesample manipulation, large preconcentration factors can generally be realisedin relatively short times with low analytical blanks.

Atomic Absorption Spectrometry

Cranston and Murray [171, 188] took the samples in polyethylene bottles thathad been precleaned at 20 ◦C for 4 days with 1% distilled hydrochloric acid. To-tal chromium (CrVI) + CrIII + Crp (particulate chromium) was coprecipitatedwith iron (II) hydroxide, and reduced chromium (CrIII + Crp) was coprecip-itated with iron (III) hydroxide. These coprecipitation steps were completedwithinminutesof sample collection tominimise storageproblems. The ironhy-droxide precipitates were filtered through 0.4 µm nucleopore filters and stored


Table 5.3. Determination of dissolved chromium species in some seawaters

Location Chromium found (µg/l)∗

Cr(III) Cr(VI) Cr bound Cr total

Port Hacking 0.27 ± 0.02 0.49 ± 0.03 0.56 ± 0.07 1.32 ± 0.05Georges River 0.42 ± 0.04 0.89 ± 0.04 0.42 ± 0.08 1.72 ± 0.06Drummoyne Bay 0.32 ± 0.03 0.95 ± 0.04 0.69 ± 0.10 1.96 ± 0.07Botany Bay 0.45 ± 0.04 1.26 ± 0.06 0.71 ± 0.03 2.41 ± 0.09Cooks River 0.51 ± 0.04 2.98 ± 0.11 0.88 ± 0.10 4.37 ± 0.06Parramatta River 0.88 ± 0.02 3.17 ± 0.06 0.82 ± 0.09 4.87 ± 0.11

∗: all results are the mean (± SD) of three measurementsFrom [189]

in polyethylene vials for later analyses in the laboratory. Particulate chromiumwas also obtained by filtering unaltered samples through 0.4 µm filters. Theiron hydroxide coprecipitates were dissolved in 6 M distilled hydrochloric acidand analysed by flameless atomic absorption. The limit of detection of thismethod is about 0.1–0.2 nmol/l. Precision is about 5%.

Mullins [189] has described a procedure for determining the concentrationsofdissolvedchromiumspecies in seawater.Chromium(III) andchromium(VI)separated by coprecipitation with hydrated iron (III) oxide and total dissolvedchromium are determined separately by conversion to chromium (VI), extrac-tion with ammonium pyrrolidine diethyl dithiocarbamate into methyl isobutylketone, and determination by atomic absorption spectroscopy. The detectionlimit is 40 ng/l Cr. The dissolved chromium not amenable to separation anddirect extraction is calculated by difference. In the waters investigated, totalconcentrations were relatively high, (1–5 µg/l), with chromium (VI) the pre-dominant species in all areas sampled with one exception, where organicallybound chromium was the major species.

A standard contact time of 4 h was found to be necessary for the quanti-tative coprecipitation of chromium on ferric oxide. The results of triplicatedeterminations of samples taken from six locations in the Sydney area arelisted in Table 5.3. The rsd values for the determinations of chromium (III),chromium (VI), and total dissolved chromium were generally 10.0%, 5.0%,and 5.0% respectively.

From these results, the rsd for the calculated concentration of the boundspecies was 20%.

X-ray Fluorescence Spectrometry

Ahern et al. [195] have discussed the speciation of chromium in seawater. Themethod used coprecipitation of trivalent and hexavalent chromium, separately,from samples of surface seawater, and determination of the chromium in the

5.15 Chromium 145

precipitates and particulate matter by thin-film X-ray fluorescence spectrome-try. An ultraviolet irradiation procedure was used to release bound metal. Theratios of labile trivalent chromium to total chromium were in the range 0.4–0.5,and the totals of labile tri- and hexavalent chromium were in the range 0.3–0.6 µg/l. Bound chromium ranged from 0 to 3 µg/l, and represented 0–90% oftotal dissolved chromium. Acidification of the samples in the usual mannerfor the determination of trace metals altered the proportion of trivalent tohexavalent chromium.

High-Performance Liquid Chromatography – Mass Spectrometry

Parts per billion concentrations of chromium (III) and chromium (VI) in sea-water have been determined using high-performance liquid chromatographyin conjunction with inductively coupled plasma mass spectrometry [196].

5.15.4Chromium (III) and Total Chromium. Gas Chromatography

Mugo and Orlans [197] have discussed shipboard methods for the determi-nation of chromium (III) and total chromium in seawater by derivatisationwith trifluoroacetylacetone followed by gas chromatography using an electroncapture detector.

5.15.5Organic Forms of Chromium

In the determination of the two oxidation states of chromium, the calculationof one oxidation state by difference presupposes that the two oxidation statesin question were statistically the only contributors to the total concentration.Because of this, contributions from other possible species such as organic com-plexes were generally not considered. It has been suggested [198] however, thatthis presumption may not be warranted, and that contributions from organ-ically bound chromium should be considered. This arises from the reportedpresence of dissolved organic species in natural waters, which form stable sol-uble complexes with chromium, and which may not be readily amenable todetermination by procedures commonly in use. The results of research intothe valency of chromium present in seawater has not always been consistent.For instance, Grimaud and Michard [199] reported that chromium (III) pre-dominates in the equatorial region of the Pacific Ocean, whereas Cranston andMurray [171] found that practically all chromium is in the hexavalent statein the northeast Pacific. Organic CrIII complexes may be formed under theconditions prevailing in seawater as well as inorganic CrIII and CrVI forms.Inconsistencies in earlier research may therefore be at least partly due to thefact that the possibility of organic chromium species was ignored [198, 200].


Nakayama et al. [201] have described a method for the determination ofchromium (III), chromium (VI), and organically bound chromium in seawater.

They found that seawater in the Sea of Japan contained about 9 ×10–9 Mdissolved chromium. This is shown to be apportioned to about 15% inorganicCrIII, about 25% inorganic CrVI, and about 60% organically bound chromium.

These workers studied the coprecipitation behaviours of chromium specieswith hydrated iron III and bismuth oxides.

The collection behaviour of chromium species was examined as follows.Seawater (400 ml) spiked with 10–8 M CrIII, CrVI, and CrIII organic complexeslabelled with 51Cr was adjusted to the desired pH by hydrochloric acid orsodium hydroxide. An appropriate amount of hydrated iron (III) or bismuthoxide was added; the oxide precipitates were prepared separately and washedthoroughly with distilled water before use [200]. After about 24 h, the sampleswere filtered on 0.4 µm nucleopore filters. The separated precipitates were dis-solved with hydrochloric acid, and the solutions thus obtained were used forγ -activity measurements. In the examination of solvent extraction, chromiumwas measured by using 51Cr, while iron and bismuth were measured by elec-trothermal atomic absorption spectrometry. The decomposition of organiccomplexes and other procedures were also examined by electrothermal atomicabsorption spectrometry.

Collection of CrIII and CrVI with Hydrated IronIII or Bismuth Oxide

Only CrIII coprecipitates quantitatively with hydrated iron (III) oxide at the pHof seawater, around 8. To collect CrVI directly without pretreatment, e.g., re-duction to CrIII, hydrated bismuth oxide, which forms an insoluble compoundwith CrVI, was used. CrIII is collected with hydrated bismuth oxide (50 mg400 ml–1 seawater). To collect CrVI in seawater a pH of about 4 was used.Both CrIII and CrVI are thus collected quantitatively at the pH of seawater,around 8.

Collection of CrIII Organic Complexes with Hydrated Iron (III) or Bismuth Oxide

The percentage collection of CrIII with hydrated iron (III) oxide may decreaseconsiderably in the neutral pH range when organic materials capable of com-bining with CrIII, such as citric acid and certain amino acids, are added tothe seawater [211]. Moreover, synthesised organic CrIII complexes are scarcelycollected with hydrated iron (III) oxide over a wide pH range [142].

As it was not known what kind of organic matter acts as the major ligand forchromium in seawater, Nakayama et al. [201] used EDTA and 8-quinolinol-5-sulfonic acid to examine the collection and decomposition of organic chromi-um species, because these ligands form quite stable water-soluble complexeswith CrIII, although they are not actually present in seawater. Both these CrIII

5.15 Chromium 147

chelates are stable in seawater at pH 8.1, and are hardly collected with either ofthe hydrated oxides. The organic chromium species were then decomposed toinorganic chromium (III) and chromium (VI) species by boiling with 1 g am-monium persulfate per 400 ml–1 seawater acidified to 0.1 M with hydrochloricacid. Iron and bismuth, which would interfere in atomic absorption spec-trometry, were 99.9% removed by extraction from 2 M hydrochloric acid solu-tion with a p-xylene solution of 5% tri-iso-octylamine. CrIII remained almostquantitatively in the aqueous phase in the concentration range 10–9 –10–6 M,whether or not iron or bismuth was present. However, as about 95% of CrVI

was extracted by the same method, samples which may contain CrVI shouldbe treated with ascorbic acid before extraction so as to reduce CrVI to CrIII.

When the residue obtained by the evaporation of the aqueous phase afterthe extraction was dissolved in 0.1 M nitric acid and the resulting solutionwas used for electrothermal atomic absorption spectroscopy, a negative inter-ference was observed, seemingly due to organic matter. This interference wassuccessfully removed by digesting the residue on a hot plate with 1 ml of con-centrated hydrochloric acid and 3 ml of concentrated nitric acid. This processhad the advantage that the interference of chloride in the atomic absorptionspectroscopy was eliminated during the heating with nitric acid.

Typically, seawater samples taken in the Sea of Japan and Pacific Oceancontained 1–1.8 ×10–9 M chromium (III), 1.7–4.5 ×10–9 M chromium (VI),3.5–6.2 ×10–9 M organic chromium, and 7.1–11.7 ×10–9 M total chromium.

Ishibashi and Shigematsu [202] used coprecipitation with aluminium hy-droxide and did not employ reduction, so that the value reported most likelycorresponds to inorganic CrIII alone; in fact, the present value for inorganicCrIII is in remarkable agreement.

In Chuecas and Riley’s study [203] the samples were stored for a longtime under acidic conditions before analysis, so that CrVI could have beenreduced to CrIII and any organic chromium dissociated, with the result thatall chromium species would have been determined as CrIII. When a sampleis reduced under acidic conditions, organic chromium is likely to dissociatepartly, initially increasing the apparent concentration of CrVI. When the an-alytical procedure of Kuwamoto and Murai [206] was re-examined, the valuefor CrIII was found actually to be the sum of CrIII and CrVI, while the valuefor CrVI was partly organic chromium. For the same reason, the CrVI valuesdetermined by Fukai [204], Fukai and Vas [205], and Yamamoto et al. [201]probably include organic chromium species. When an iron (II) precipitate isused, there seems to be little chance of determining the organic chromiumspecies as CrVI. The value for CrVI reported by Cranston and Murray [171]agrees quite well with the value for CrVI reported by them, although the valuefor CrIII is lower. The results obtained by Grimaud and Michard [199] for CrIII

differ considerably, but the discrepancies cannot be discussed because detailsof the analytical procedure were not given. It seems reasonable to conclude thatthe inconsistency of past results concerning the dominant chromium species


and the total chromium concentration in seawater can be attributed, at leastin part, to the fact that the presence of organic chromium species was notconsidered properly.

See also Sects. 5.74.5, 5.74.6, 5.74.8–5.74.12, and 5.74.14–5.74.16.

5.16Cobalt

Little is known of the oceanic distribution or speciation of cobalt, because verylow concentrations (< 200 pM) make its determination difficult. Laboratorystudies indicate that cobalt exists in seawater primarily as the cobalt (II) ion andas the carbonate complex. Organic complexes are not considered important.

The oceanic distribution of cobalt is similar to that of manganese, althoughcobalt concentrations are 10–100 times smaller; maximum concentrations are100–300 pM in surface waters, decreasing to 10 pM at depths below 1000 m. Asconcentrations of cobalt in seawater are so low, it may become biolimiting inopen ocean surface waters.

Because cobalt is an essential element in biological compounds like vitaminB12 and some metalloproteins [208,209], the low concentration of this metal inseawater points to the possible role of cobalt as a biolimiting nutrient [210,211].The discharge of various cobalt radionucleides from nuclear installations tocoastal waters and their accumulation by marine organisms [212–214] has alsoincreased interest in the fate of this element. Dissolved cobalt occurs in seawa-ter at concentrations ranging from 0.01 to 0.2 nM [210,211]. Calculation of theorganic complexation of cobalt using an ion pairing model and stability con-stants valid for seawater [215] shows that it is weakly complexed by inorganicligands, the predominant inorganic species being cobalt (II) and as chloridecomplexes. There is some evidence that cobalt in seawater occurs stronglycomplexed by organic ligands [216, 217]. The available data on cobalt distri-bution in seawater [210, 218–220] show surface minima, a maximum withinthe upper thermocline as a result of atmospheric input, and depletion at depthdue to removal from seawater, probably in association with MnO [221, 222].


Various methods have been proposed for the determination of traces of cobaltin seawaterandbrines,mostnecessitatingpreconcentration.Solventextractionfollowed by spectrophotometric measurements [223–230] is the most popularmethod but has many sources of error; the big difference in the volumes ofthe two phases results in mixing difficulties, and the solubility of the organicsolvent in the aqueous phase changes the volume of organic phase, resultingin decreased reproducibility of the measurements. In many cases, excess of

5.16 Cobalt 149

reagent and various metal complexes are co-extracted with cobalt and causeerrors in determining the absorbance of the cobalt complex.

The procedure of Kentner and Zeitlin [223] is as follows: to a filtered 750 mlsample of seawater add 20% aqueous sodium citrate (25 ml), 30% aqueous hy-drogen peroxide (1 ml), and 1% ethanolic l-nitroso-2-naphthol (treated withactivated charcoal and filtered) (1 ml), and set aside for 10 min. Extract thecobalt complex into chloroform and back-extract the excess of reagent intobasic wash solution (mix 1 M sodium hydroxide (50 ml), 20% aqueous sodiumcitrate (10 ml), and 30% aqueous hydrogen peroxide (1 ml) with water to pro-duce 100 ml (3×5 ml extractions), shaking for 60 s for each extraction). Extractcopper and nickel from the organic phase into 2 M hydrochloric acid (5 ml),back-extract any released reagent into basic wash solution (5 ml), and washthe chloroform phase again with 2 M hydrochloric acid (5 ml). Dilute the or-ganic phase to 50 ml (for 30 ml cells) or 30 ml (for 20 ml cells) and measure theextinction at 410 nm against a blank in 10 cm cells.

In another spectrophotometric procedure Motomizu [224] adds to the sam-ple (2 litres) 40% (w/v) sodium citrate dihydrate solution (10 ml) and a 0.2%solution of 2-ethylamino-5-nitrosophenol in 0.01 M hydrochloric acid (20 ml).After 30 min, add 10% aqueous EDTA (10 ml) and 1,2-dichloroethane (20 ml),mechanically shake the mixture for 10 minutes, separate the organic phaseand wash it successively with hydrochloric acid (1:2) (3 × 5 ml), potassiumhydroxide (5 ml), and hydrochloric acid (1:2) (5 ml). Filter, and measure theextinction at 462 nm in a 50 mm cell. Determine the reagent blank by addingEDTA solution before the citrate solution. The sample is either set aside forabout 1 day before analysis (the organic extract should then be centrifuged),or preferably it is passed through a 0.45 µm membrane-filter. The optimum pHrange for samples is 5.5–7.5. From 0.07 to 0.12 µg/l of cobalt was determined;there is no interference from species commonly present in seawater.

Kouimtzis et al. [232] described a spectrophotometric method for down toµg/l cobalt in seawater, in which the cobalt is extracted with 2,2′-dipyridyl-2-pyridylhydrazone (DPPH) [233,235–238], the cobalt complex is back-extractedinto 20% perchloric acid, and this solution is evaluated spectrophotometricallyat 500 nm. This avoids many of the sources of error that occur in earlierprocedures.

Isshiki and Nakayama [234] have discussed the selective concentration ofcobalt in seawater by complexation with various ligands or sorption on macro-porous XADresins. Complexed cobalt is collected after passage through a smallXAD resin-packed column.


Analytical procedures for the determination of cobalt in seawater generallyuse graphite furnace absorption spectrometry after a preconcentration step


involving solvent extraction, coprecipitation, or ion exchange on Chelex 100resin [239].These techniqueshavedifficulties achieving the sensitivity requiredfor the determination of the low levels of cobalt in seawater, and include a highrisk of sample contamination and loss of analyte during the several samplepreparation steps involved.


Malahoff et al. [240] used a shipboard flow injection spectrophotometric tech-nique to determine ppt concentrations of cobalt in seawater

5.16.4Atomic Fluorescence Spectrometry

Yuzefovsky et al. [241] used C18 resin to preconcentrate cobalt from seawaterprior to determination at the ppt level by laser-excited atomic fluorescencespectrometry with graphite electrothermal atomiser.

5.16.5Spectrofluorometry

To determine cobalt in seawater, Boyle [242] solvent-extracted the cobalt usingLuminol, with which 5 pmol kg–1 of cobalt can be determined.

5.16.6Chemical Luminescence Analysis

Sakamoto [243] determined picomolar levels of cobalt in seawater by flowinjection analysis with chemiluminescence detection. In this method flow in-jection analysis was used to automate the determination of cobalt in seawaterby the cobalt-enhanced chemiluminescence oxidation of gallic acid in alkalinehydrogen peroxide. A preconcentration/separation step in the flow injectionanalysis manifold with an in-line column of immobilised 8-hydroxyquinolinewas included to separate the cobalt from alkaline-earth ions. One sample anal-ysis takes 8 min, including the 4-min sample load period. The detection limitis approximately 8 pM. The average standard deviation of replicate analysesat sea of 80 samples was ±5%. The method was tested and intercalibrated onsamples collected off the California coast.

Cobalt (II) has been determined by online measurements on seawater whichhas been passed through a column containing 8-quinolinol immobilised onsilica gel, followed by chemical luminescence detection [244].

5.16 Cobalt 151


Adsorptive cathodic stripping voltammetry has an advantage over graphitefurnace atomic absorption spectrometry in that the metal preconcentrationis performed in situ, hence reducing analysis time and risk of contamina-tion. Additional advantages are low cost of instrumentation and maintenance,and the possibility to use adapted instrumentation for online and shipboardmonitoring.

Vega and Van den Berg [245] have described a procedure for the directdetermination of picomolar levels of cobalt in seawater. Cathodic strippingvoltammetry is preceded by adsorptive accumulation of the cobalt–nioxime(cyclohexane-1.2-dione dioxime) complex from seawater containing 6 µMnioxime and 80 mM ammonia at pH 9.1 onto a hanging mercurydrop elec-trode, followed by reduction of the adsorbed species. The reduction currentis catalytically enhanced by the presence of 0.5 M nitrite. Optimised condi-tions for cobalt include a 30 s adsorption period at –0.7 V and a voltammetricscan using differential pulse modulation. According to the proposed reactionmechanism, dissolved cobalt (II) is oxidised to cobalt (III) upon addition ofnioxime and high concentrations of ammonia and nitrite; a mixed cobalt (III)–ammonia–nitrite complex is adsorbed on the electrode surface; the cobalt (III)is reduced to cobalt (II) (complexed by nioxime) during the voltammetric scan,followed by its chemical reoxidation by the nitrite, initiating a catalytically en-hanced current. A detection limit of 3 pM cobalt (at an adsorption period of60 s) enables the detection of this metal in uncontaminated seawater usinga very short adsorption time. UV digestion of seawater is essential, as partof the cobalt may occur strongly complexed by organic matter and renderednonlabile. The method was applied successfully to the determination of thedistribution of cobalt in the water column of the Mediterranean.

See also Sects. 5.74.4–5.74.6 and 5.74.8–5.74.16.

5.16.8Polarography

Harvey and Dutton [231] determined nanogram amounts of cobalt in seawaterafter concentration on manganese dioxide formed by photochemical oxida-tion of divalent manganese in a photochemical reactor. The sample (1 litre)containing 100 µg manganese was irradiated in an Hanovia reactor fitted witha 2 W low-pressure Hg discharge lamp radiating mainly at 254 and 185 nm.The manganese dioxide deposit that adhered to the silica jacket of the reactorwas dissolved in 0.15 M hydrochloric acid containing a trace of sulfur dioxide,the solution was evaporated to dryness, and the residue was dissolved in 4 mlof 0.625 M hydrochloric acid; 1 ml 5 M aqueous ammonia and 0.1 ml of 0.1%dimethylgloxime in ethanol were added, and cobalt was determined by pulse


polarography. The polarograph was operated in the derivative mode, startingat –1.0 V, and a 50 mV pulse height and 1 s mercury drop life were used.

5.17Copper

Copper (II) is present in natural waters in a variety of chemical forms. Syl-va [246] indicated that the following species are found in freshwater systems:Cu2+; CuCO3; Cu(CO3)2–

2 ; CuHCO+3 ; CuOH+; Cu2(OH)2+

2 ; CuCl+. It was alsofound that Cu2+ can be removed completely from aquatic systems by precipi-tation as Cu(OH)2, CuCO3, and Cu(OH)n(CO3)1–n/2.

Sunda and Hanson [247] have used ligand competition techniques for theanalysis of free copper (II) in seawater. This work demonstrated that only 0.02–2% of dissolved copper (II) is accounted for by inorganic species. (i.e., Cu2+,CuCO3, Cu(OH)+, CuCl+, etc.); the remainder is associated with organic com-plexes. Clearly, the speciation of copper (II) in seawater is markedly differentfrom that in fresh water.

Importantly, Sunda and coworkers [248, 249] demonstrated that free cop-per (II) – not total copper – is responsible for copper (II) toxicity. Consequently,the impact of copper (II) on the marine environment can be ascertained onlyby measurement of free copper (II) levels.

The importance of complexing agents in the mineral nutrition of phyto-plankton and other marine organisms has been recognised for more than20 years. Complexing agents have been held responsible for the solubilisationof iron, and therefore its greater biological availability [250]. In contrast, com-plexing agents are assumed to reduce the biological availability of copper andminimise its toxic effect [251–264]. Experiments with pure cultures of phyto-plankton in chemically defined media have demonstrated that copper toxicityis directly correlated with cupric ion activity and independent of the totalcopper concentration. In these experiments, cupric ion (Cu2+) concentrationswere varied in media containing a wide range of total concentrations throughthe use of artificial complexing agents. When Cu2+ concentration was calcu-lated for earlier experiments with phytoplankton in defined media, it appearedthat Cu2+ was toxic to a number of phytoplankton species in concentrations aslow as 10–6 µmol–1 [260]. Since copper concentrations in the world ocean typ-ically range from 10–4 to 10–1 µmol–1, complexing agents and other materialsaffecting the solution chemistry of copper must maintain the Cu2+ activity atsublethal levels.

Copper may exist in particulate, colloidal, and dissolved forms in seawater.In theabsenceoforganic ligands,orparticulate andcolloidal species, carbonateand hydroxide complexes account for more than 98% of the inorganic copperin seawater [285, 286]. The Cu2+ concentration can be calculated if pH, ionicstrength, and the necessary stability constants are known [215, 265–267]. Inmost natural systems, the presence of organic materials and sorptive surfaces

5.17 Copper 153

significantly alters speciation and decreases the utility of equilibrium calcula-tions. Analytical difficulties in the measurement of Cu2+ and copper associatedwith naturally occurring ligands has encouraged numerous workers to intro-duce the “complexation capacity” concept [268–270]. Functionally, the coppercomplexing capacity of a water sample is the ability of the sample to removeadded copper from the free ion pool [271]. Analytically, complexation capacitymeasurements depend on quantitation complexing ability of an operationallydefined group of ligands. The assumption is made that unknown ligands maybe classed into meaningful groups on the basis of the physical properties oftheir metallo-complexes, (e.g., lability to anodic stripping voltammetry (ASV),chelating resins, or ultraviolet radiation). Schemes to determine the concen-tration of copper associated with different classes have been proposed as usefulways to address complexing capacity questions in natural systems [272,273], ashave various titrametric techniques [274,275]. Different analytical proceduresmeasure the copper chelating capacity of slightly different classes of ligands,and there is some overlap in the complexes included in classes defined by dif-ferent techniques. For example, while there is a small fraction of organic ma-terial in seawater which forms ASV-labile complexes not dissociated by Chelexresin [276], most AVS-labile complexes are also labile to chelating resins [277].

5.17.1Titration Procedures

Ruzic [278] considered the theoretical aspects of thedirect titrationof copper inseawaters and the information this technique provides regarding copper speci-ation. The method is based on a graph of the ratio between the free and boundmetal concentration versus the free metal concentration. The application ofthis method, which is based on a 1:1 complex formation model, is discussedwith respect to trace metal speciation in natural waters. Procedures for inter-pretation of experimental results are proposed for those cases in which twotypes of complexes with different conditional stability constants are formed,or om which the metal is adsorbed on colloidal particles. The advantages of themethod in comparison with earlier methods are presented theoretically andillustrated with some experiments on copper (II) in seawater. The limitationsof the method are also discussed.

Waite and Morel [279] have described an amperometric titration procedurefor the characterisation of organic copper complexing ligands, and appliedit to a variety of synthetic and naturally occurring organic compounds. Theprocedure is based upon the ability, in solutions of high chloride content, toobtain a sensitive and reproducible amperometric measurement of reduciblecopper (II) at positive voltages up to about 100 mV relative to an Ag/AgCl refer-ence electrode. Copper (II) is reduced to copper (I), which is stabilised by chlo-ride despite the presence of oxygen. Application of the titration technique toahigh chloride content electrolyte containingvarious concentrationsofnitrilo-


triacetic acid confirms that copper–ligand reduction and dissociation are notmajor problems, provided that a sufficiently positive working electrode poten-tial is chosenand that the concentrationof theorganic ligand is low.Applicationof the procedure to a variety of naturally occurring organic agents includinga fulvic acid, fresh water algal exudates, and a sample of Sargasso seawater,produces results that are consistent with those found by alternative methods.


Cabon and Le Bihan [280] studied signals obtained by electrothermal atomicabsorption spectrometry of seawater matrices. The interference effect of sodi-um chloride, sodium nitrate, sodium sulfate, magnesium chloride, and calciumchloride on the electrothermal atomic absorption spectrometric signal of cop-per in seawaterwere studied.Thermal treatmentat 600–800 ◦Ccausedhydroly-sis of magnesium chloride to magnesium oxide and minimised its interference.Ashing at higher temperatures of 1300, 1200, and 1100 ◦C was carried out in thepresence of sulfate, nitrate, and chloride salts, respectively, without loss of cop-per. A study of the influence of two-component, chloride–nitrate or chloride–sulfate matrices illustrated the stabilising effect of the formation of metaloxides and metal sulfides on the copper signal. This stabilisation enhanced thedecrease in interference connected with chloride removal in an acidic medium.

In further work Cabon [281] proposed a model to describe the variationsin copper signals caused by the principal inorganic ions in seawater (sodium,magnesium, calcium, chloride, and sulfate). Data obtained by ashing simu-lated seawaters under different temperature conditions were used. Ashing at800 ◦C caused hydrolysis of magnesium chloride to magnesium oxide and theformation of sodium sulfide. Both of these products enhanced the stability ofcopper in the furnace. A complementary decrease in interference occurred inthe presence of magnesium when a small amount of nitrate (0.2 M) was added.The model was confirmed by results obtained using nitric or sulfuric acid asmodifier.

To determine down to 2.4 µmol/l of copper in seawater, Nishoika et al. [282]complexed the copper with di-ethyl-dithio carbamate, precipitated with fer-ric hydroxide, filtered off and dissolved the precipitate with nitric acid, anddetermined copper by electrothermal atomic absorption spectrometry.

Muzzarelli and Rocchetti [287] showed that chitosan is superior to DowexA700 ion exchange resin and modified celluloses for the collection of copperfrom unoxidised seawater. In this procedure, the sample is passed througha column (30 × 3 mm) packed with chitosan (100 mg: 100–200 mesh), andthe chelated copper eluted with a 1% solution of 1.10-phenanthroline (20 ml)at 50 ◦C or with 1 M sulfuric acid (20 ml). Place an aliquot (20 µl) in a hotgraphite analyser programmed to dry for 20 s, char for 20 s and atomise for10 s. Determine the amount of copper present by comparison with standards.

5.17 Copper 155

Average recoveries from the column were 100%, and the coefficient of variationwas 7.5% for 3.4 µg of copper per litre.

5.17.3Spectrophotometric Method and Spectrofluorometric Method

Abraham et al. [286] determined total copper in seawater spectrophotometri-cally using quinaldehyde 2-quinolyl-hydrazone as chromogenic reagent. Thismethod is capable of determining copper at the ppb level.

Hui-Hui Zeng et al. have described a real-time determination of free cop-per (II) in seawater using a fluorescence-based fibre optic method [946].

A C18 column loaded with sodium diethyldithiocarbamate has been usedto extract copper and cadmium from seawater. Detection limits for analy-sis by graphite furnace atomic absorption spectrometry were 0.024 µg/l and0.004 µg/l, respectively [283].


Ion-selective electrodes have been used for the potentiometric determinationof the total cupric ion content of seawater [284]. Down to 2 µg/l cupric coppercould be determined by this procedure.

De Marco [285] evaluated three different types of copper (II) ion-selectiveelectrodes. The copper sulfide electrode was found to be oxidised, whereas thecopper selenide and copper sulfide electrodes were found to have chloride ioninterference at copper (II) activities exceeding 10–8 M.

Prior to the introduction of ion-selective electrode techniques, in situ mon-itoring of free copper (II) in seawater was not possible due to the practicallimitations of existing techniques (e.g., ligand competition and bacterial reac-tions). Ex situ analysis of free copper (II) is prone to experimental error, asthe removal of seawater from the ocean can lead to speciation of copper (II).Potentially, a copper (II) ion electrode is capable of rapid in situ monitoring ofenvironmental free copper (II). Unfortunately, copper (II) has not been usedwidely for the analysis of seawater due to chloride interference that is allegedto render the copper nonfunctional in this matrix [288].

Westall et al. [289] and Lewenstam et al. [290] proposed a layer mechanismto account for the chloride interference in the CuS electrode.

5.17.5Electroanalytical Methods

Cathodic stripping voltammetry has been used to determine copper species inseawater [291, 292].


In one method [292], a reduction current occurred when a solution of cate-chol and copper was subjected to cathodic stripping voltammetry at a hangingmercury drop electrode. The composition of the adsorbed film and the opti-mal conditions for its formation were investigated. The phenomenon was usedto determine copper in seawater using ac polarography to measure complexadsorption. Currents were detected at the very low copper concentrations thatoccurred in uncontaminated seawater. Competition for copper ions by natu-ral organic complexing ligands was evident at low concentrations of catechol.The method was more sensitive and had a shorter collection period than therotating-disk electrode DPASV technique, with comparable accuracy.

Scorano et al. [293] determined copper at the 5 ×10–10 M level in seawaterby anodic stripping voltammetry using ethylenediamine. These workers inves-tigated the properties of ethylenediamine (en) as a means of performing theanalysis of ligand-exchangeable and labile (i.e., directly reducible at pH 8) cop-per in seawater at trace levels. Stripping polarographic or pseudopolarographicdeterminations show that the copper ethylenediamine complex behaves re-versibly in seawater, exchanging two electrons at the mercury drop electrode.

The role of chloride ions in competitive reactions with ethylenediamine forcopper during the stripping step was also studied. In seawater made 2 ×10–3 Min ethylenediamine, copper (II) can be detected at the hanging mercury dropelectrode by differential pulse anodic stripping voltammetry at the 5 ×10–10 Mlevel, with a deposition time of 10 min. A procedure for measuring pH 8 labilecopper in seawater is obtained by coupling differential pulse anodic strippingvoltammetry with a medium alteration method. Addition of ethylenediamineat the end of the electrolysis more than doubles peak height by doubling thecurrent yield per mole of copper and by removing interferences associatedwith the oxidation of copper in chloride media. This procedure facilitates thevoltammetric study of copper in seawater under natural conditions.

Quentel et al. [294] complexed copper with 1,2-dihydroxyanthraquinone-3sulfuric acid prior to determination by absorptive stripping voltammetry inamounts down to 0.3 nM in seawater.

Wang et al. [295] used a remote electrode, operated in the potentiometricstripping mode, for the continuous onboard measurement of copper distribu-tion patterns in San Diego Bay (CA, USA).

Garcia-Monco Carrá et al. [296] have described a “hybrid” mercury filmelectrode for the voltammetric analysis of copper (and lead) in acidified sea-water. Mercury plating conditions for preparing a consistently reproduciblemercury film electrode on a glassy carbon substrate in acid media are evalu-ated. It is found that a “hybrid electrode”, i.e., one preplated with mercury andthen replated with mercury in situ with the sample, gives very reproducible re-sults in the analysis of copper in seawater. Consistently reproducible electrodeperformance allows for the calculation of a cell constant and prediction of theslopes of standard addition plots, useful parameters in the study of copperspeciation in seawater.

5.17 Copper 157

A hanging mercury drop electrodeposition technique has been used [297]for a carbon filament flameless atomic absorption spectrometric method forthe determination of copper in seawater. In this method, copper is transferredto the mercury drop in a simple three-electrode cell (including a counter-electrode)byelectrolysis for30 minat–0.35 Vversus theSCE.Afterelectrolysis,the drop is rinsed and transferred directly to a prepositioned water-cooledcarbon-filamentatomiser, and themercury is volatilisedbyheating thefilamentto 425 ◦C. Copper is then atomised and determined by atomic absorption. Thedetection limit is 0.2 µg copper per litre simulated seawater.

5.17.6Isotope Dilution Methods

Isotope dilution mass spectrometry has been used to determine traces ofcopper in seawater [298, 299].

5.17.7Electron Spin Resonance Spectrometry

Background copper levels in seawater have been measured by electron spinresonance techniques [300]. The copper was extracted from the seawater intoa solutionof8-hydroxyquinoline inethylpropionate (3 mlextractantper100 mlseawater), and the organic phase (1 ml) was introduced into the electron spinresonance tube for analysis. Signal-to-noise ratio was very good for the four-line spectrum of the sample and of the sample spiked with 4 and 8 ng Cu2+.The graph of signal intensity versus concentration of copper was rectilinearover the range 2–10 µg/l of seawater, and the coefficient of variation was 3%.

5.17.8Miscellaneous Methods

Traces of copper and lead are separated [301] from macro amounts of calcium,magnesium, sodium, and potassium by adsorption from the sample onto activecarbon modified with hydroxyquinoline dithizone or diethyldithiocarbamate.

Marvin et al. [302] have discussed the effects of sample filtration on thedetermination of copper in seawater, and concluded that glass filters couldseriously affect the reliability of subsequent analysis.

5.17.9Copper Speciation

Wood et al. [303] have described an ion exchange technique for the mea-surement of the copper complexing capacity of seawater samples taken in the


Sargasso Sea, and continental shelf samples. This technique measures the cop-per complexing capacity of relatively strong dissolved and colloidal organiccomplexing agents in natural seawater. The technique was used to compare thecopper complexing capacity of strong organic dissolved and colloidal complex-ing agents in those samples. They also determined the relationship betweenthe copper complexing capacity of a specific group of complexing agents andthe concentration of two large heterogeneous pools of potential complexingagents; dissolved organic carbon and total particulate material. The coppercomplexing capacity of these samples ranged from 0.014 to 1.681 µmol Cuper litre on the inner shelf, from 0.043 to 0.095 mol Cu per litre in mid- andouter-shelf waters, and from 0.010 to 0.036 µmol Cu per litre at the SargassoSea stations.

The ion exchange procedure used by Wood et al. [303] to estimate cop-per complexation capacity was a modification of that used by Stolzberg andRosin [304] and Giesy [305]. Excess Cu2+ is added to the filtered samples and al-lowed to equilibrate with available ligands. Each sample is then passed througha column packed with Nahelex resin (Biorad 100–200 mesh), and Cu2+ andCu associated with weak or rapidly dissociating complexes are removed by theresin; Cu remaining in the sample after chromatography provides a quantita-tive measure of the Cu-chelating capacity of strong ligands remaining in thesample. The procedure has the advantage that complex formation proceedsat seawater pH in a relatively undisturbed sample. However, the procedurealso depends on the assumption that essentially all the Cu ligands (Ln–) in thesample are associated with Cu.

This involves the reaction

mCu2+ + Ln– = CumL(n–2m)–

All chromatography was conducted at flow rates greater than 20 ml cm–2 s–1,since slower flow rates resulted in complex dissociation.

The concentration of copper in the column eluent was determined by flameatomic absorption spectroscopy of samples which were preconcentrated withammonium pyrrolidine dithiocarbamate (APDC) and methyl isobutyl ketone.The pH of the acidified sample was adjusted to pH 2.5–3.5 using 400 µl 8 Mammonium acetate (Chelex cleaned).

Zorkin et al. [306] developed a procedure to estimate the amount of bio-logically active copper in seawater based on the assumption that the divalentcopper ion was the most toxic species, and its concentration could be relatedto the amount of metal sorbed on a sulfuric-acid cation exchange resin. Themethod was tested by application to artificial seawater containing copper andthe organic ligands EDTA, NTA, histidine, and glutamic acid. Other experi-ments showed there was a correlation between the inorganic copper fractiondetermined by the ion exchange procedure and the toxic fraction of copperquantified by a bioassay using the marine diatom.

5.17 Copper 159

In recent years much of the work concerned with the speciation of cop-per and other trace metals in natural waters has been done using anodicstripping voltammetry. This work has primarily progressed in two directions:studies of the shift in trace metal peak potentials with changing concentra-tions of ligands [307–311], and studies of changes in metal peak height orpeak area under differing experimental conditions. Variants of the second ap-proach include pH titrations [311–313] and compleximetric titrations [269],in which natural or added ligands are quantitatively titrated with metal ions,or alternatively metal ions are titrated with ligands [269, 270, 314]. In thistechnique, the electrolysis potential is set at a value which presumably dis-criminates between the “free” (i.e., rapidly reducible) metal and the com-plexed metal, which is reduced at a much slower rate. This approach hasbeen used extensively to estimate the “complexation capacity” of naturalwaters.

Techniques based on the shift of the peak potential depend on the degreeof reactivity of the oxidised metal with the ligand of interest in the reactionlayer. They can describe the species undergoing reduction, i.e., the speciationin the natural medium, only indirectly and by assuming reversibility. Thusthey are more suitable for model studies [265, 307] and for the determinationof stability constants in known media [308, 309, 315] than for direct deter-mination of natural speciation. On the other hand, methods dependent onpeak height or peak area can give direct information on the natural species aslong as a direct proportionality exists between the quantity of metal reducedduring electrolysis and the metal oxidised from the amalgam. One relativelynovel form of anodic stripping voltammetry which gives information aboutthe species undergoing reduction is stripping polarography, sometimes calledpseudopolarography [316–318].

In this technique, peak heights or peak areas obtained by anodic strippingvoltammetry are plotted against the applied electrolysis potential. These plotshave the sigmoidal shape of ordinary dc polarograms, but without the residualcurrent component, and present the possibility of extending existing polaro-graphic methodology to trace metals at the part per billion (µg/l) level. Theshapes of the plots indicate the degree of reversibility of the species undergoingreduction and may be useful for their identification.

Zirino and Kounaves [319] applied this technique to a study of the reductionof copper in seawater. Figure 5.8 shows three plots of 6 µg/l copper added tounfiltered seawater from San Diego Bay (CA, USA), and analysed under varyingconditions. Each of the experimental points represents the copper peak currentobtained by anodic stripping voltammetry after 5 min electrolysis at a hangingmercury drop electrode. The plots obtained for copper at pH 8 (Figs. 5.8aand 5.8b) feature broadly curving slopes, and no distinct limiting plateau isreached, even at the highest applied potential. The shapes of these “waves”indicate an irreversible reduction. On the other hand, the reduction of copperat pH 3.0 is quasi-reversible, with E3/4 – E1/4 = 42 mV (Fig. 5.8c).


Figure 5.8. Stripping polarographic plots of 6 ppb Cu in seawater (not to scale). (a) Rawseawater, pH 8. (b) Ultraviolet-oxidised seawater, pH 8. (c) Raw seawater, pH 3. Source: [319]

Potentiometric stripping analysis has been applied by Sheffrin and Wil-liams [320] to the measurement of copper in seawater at environmental pH.The advantage of this technique is that it can be used to specifically measurethe biologically active labile copper species in seawater samples at desired pHvalues. The method was applied to seawater samples that had passed a 0.45 µmMillipore filter. Samples were studied both at high and at low pH values.

These workers used a radiometer ISS 820 ion-scanning system [101, 321–323] equipped with a glassy carbon electrode to determine copper at the 2–200 µg/l level in nitrogen-purged 0.45 µm Millipore-filtered seawater to whichhad been added 5 ppm mercury.

The speciation of copper is different at high and low pH. At pH 1.0 most ofthe copper will be labile and a total copper concentration will be measured. AtpH 7.7 there should be a smaller proportion of labile copper, as much will becomplexed in various forms, depending on the constituents of the seawater.

Because of this complexation capacity, any standard addition performedat high pH will not return 100% of the spike, so a true value for the copperconcentration cannot be calculated. Therefore, after an initial measurement athigh pH the sample was acidified to pH 1.0 with 0.5 ml acid and another traceobtained. This compared the amount of copper released at low pH with thelabile fraction at high pH. Standard additions were performed on the sampleat low pH so almost all of the spike was returned. This allowed an estimate tobe made of the percentage of total copper that was labile at high pH, and thequantification of this fraction in µg/l. This is illustrated graphically in Fig. 5.9.

5.17 Copper 161

Figure 5.9. Measurements of labile copper at high pH. Source: Author’s own files

The analysis of total copper by potentiometric stripping analysis dependson the way any bound copper is released; that is, whether the sample is acidifiedand to what pH, whether it is acidified and boiled, or treated with ultravioletradiation and then acidified. Results obtained using these different meth-ods are given in Table 5.4, which compares the analysis of a “poor-quality”sample, i.e., low-pH aquarium seawater, by potentiometric stripping analy-sis, and by atomic absorption spectrometry after extraction with ammoniumtetramethylene-dithiocarbamate-isobutyl methyl ketone [324, 325], as well asa potentiometric stripping analysis of a “good-quality” sample, i.e., seawater ofhigher pH. The difference between the results for the “poor-quality” seawateranalysed by the two techniques was not significant.

Shuman and Michael [326, 327] introduced a technique that has sufficientsensitivity for kinetic measurement at very dilute solutions. It combines anodicscanning voltammetry with the rotating-disk electrode and provides a methodfor measuring kinetic dissociation rates in situ, along with a method for distin-guishing labile and non-labile complexes kinetically, consistent with the waythey are defined.

Odier and Plichon [328] used ac polarography to determine the chemicalform and concentration of copper in seawater. The shift of the E1/2 of reductionof CuII determined by ac polarography serves to identify the inorganic com-plexes of copper and to determine their formation constants. They showed thatcopper is present in seawater mainly as Cu2+, CuCl+, and (Cu(HCO3)2(OH))–.Copper down to 3 µg/l was determined in seawater by ac polarography afteracidifying to pH 5 by passage of carbon dioxide.

Turner et al. [331] studied the application of the automated electrochemicalstripper to the determination of copper in seawater.


Table 5.4. Analysis of ‘poor quality’ seawater by potentiometric stripping analysis andatomic absorption spectroscopy and ‘good quality’ seawater by potentiometric strippinganalysis

Analysis Measured pH Sample Electrolysis Plating timecopper treatment voltage (min)concentration(ppb)

Potentiometric 8.3 1.7 Acidified – 0.6 8stripping 8.6 1.1 Acidified – 1.1 8analysis of 6.8 1.7 Acidified – 0.6 8‘poor quality’ 9.6 1.7 Acidified – 0.6 16seawater 8.6 1.3 Acidified and

boiled (15 min)– 0.6 32

14.0 1.6 Acidified andboiled (15 min)

– 0.6 8


– 0.6 8

Mean (±SD) =10.24 ± 3.3

AAS/ATDC – 35 sample:IBMK∗ Mean (±SD) =analysis of 12.4 ± 3.1‘poor quality’seawater.


– 0.6 32

Potentiometric 2.2 0.9 Acidified – 0.6 32stripping 8.6 1.3 Acidified – 0.6 32analysis of‘good quality’seawater

0.36 7.0 Ultravioletirradiation(67 min)

– 0.6 32

∗: Ammonium tetramethylenedithiocarbamate – isobutyl methyl ketoneFrom Author’s own files

Neutron activation analysis has been used [329] to determine total copperin seawater.

Cathodic stripping voltammetry has been used to determine copper speciesin seawater [291, 292]. Van den Berg [330] determined copper in seawater bycathodic stripping voltammetry of complexes with catechol.

A technique for the determination of free cupric ions in seawater has beendescribed by Sunda and Hanson [332]. The method is based on sorption ofcopper onto Sep-Pak C18 cartridges and internal free cupric ion calibration.Calibration is accomplished by adding cupric ion buffers and EDTA, whichcompetes with natural organic ligands for copper complexation. The methodwas used to establish that 0–2% of the copper occurs as inorganic species and98–100% occurs as organic complexes.

See also Sects. 5.74.4–5.74.6 and 5.74.8–5.74.16.

5.18 Dysprosium 163

5.18Dysprosium

See Sect. 5.49.

5.19Erbium

See Sect. 5.49.

5.20Europium

See Sects. 5.49 and 5.74.16.

5.21Gadolinium

See Sect. 5.49.

5.22Gallium

See Sect. 5.74.14.

5.23Germanium

5.23.1Hydride Generation Furnace Atomic Absorption Spectrometry

Andreae and Froelich [333] have described a procedure for the determinationof germanium in seawater. In this method the peak absorbance was somewhatdependent upon atomisation temperature, rising sharply between 2400 and2500 ◦C, and remaining almost constant above this temperature. As the life-time of the tube decreases with the burn temperature, it was decided to use2600 ◦C as the analysis temperature. The addition of a short high-temperature(2900 ◦C) burn cycle with full-purge gas flow preceding the analysis burn cycleimproved the blank values and removed memory effects which were some-times encountered when going from large to small analyte amounts. Underthese conditions, tubes lasted for at least 100 determinations.

The sensitivity of this system is 430 pg/0.0044 absolute. The standard devi-ation of the baseline noise is about 0.0007 absolute, resulting in a noise-limiteddetection limit of 140 pg of germanium at the 95% confidence level.



5.24Gold


Falkner and Edmond [334] determined gold at femtomolar quantities in sea-water by flow injection inductively coupled plasma quadrupole mass spec-trometry. The technique involves preconcentration by anion exchange of goldas a cyanide complex, [Au(CN)–

2], using 195Au radiotracer (t1/2 = 183 days)to monitor recoveries. Samples are then introduced by flow injection into aninductively coupled plasma quadrupole mass spectrometer for analysis. Themethod has a detection limit of ≈10 fM for 4 litres of seawater preconcentratedto 1 ml, and a relative precision of 15% at the 100 fM level.

5.24.2Photometry

Pilipenko and Pavlova [335] determined traces of gold in seawater usinga “spot” photometric method. This method is based on the catalysis (byAuCl2SO4) of the oxidation of FeII by AgI with production of metallic silver.The sample is filtered through paper, and the paper is dried and decomposedwith sulfuric acid, nitric acid, hydrofluoric acid, water solution (1:1:1:1). Theresidue is dissolved in a few drops of aqua regia, this solution is evaporated,and the residue is dissolved in one drop of 0.05 M sulfuric acid. This solution isapplied to filter paper, and to the resulting spot are applied drops of phosphate–citrate buffer solution (pH 2.4) of 0.72 M ferrous sulfate, in 0.05 M sulfuric acidand of 0.1 M silver nitrate; 15 s after the silver nitrate solution has been applied,the reflectance of the spot (due to metallic silver) is measured with a suitableinstrument. The reflectance is proportional to the amount of gold on the paper,from 3 to 60 pg. As little as 0.0025 µg/l of gold can be determined in seawater.

5.25Holmium

See Sect. 5.49.

5.26Indium


Matthews and Riley [336] determined indium in seawater at concentrationsdown to 1 ng/l. Preconcentration of metals on a cation exchanger was fol-

5.27 Iridium 165

lowed by separation of alkali metals and alkaline-earth metals by retentionof indium as a chloro-complex on an anion exchanger. Samples of indiumcontaining eluate were then concentrated and irradiated in a thermal neutronflux of 5 ×1012 neutrons/cm2/s for several weeks, and the resulting 1.98 MeVβ-radiation of the long-lived 114mln daughter nuclide was counted. Minor ele-ments were removed by a series of post-irradiation solvent-extraction stages.

See also Sects. 5.74.5, 5.74.8, and 5.74.9.

5.27Iridium

See Sect. 5.74.5.

5.28Iron

Worldwide, marine chemists and marine biologists have focused on the be-haviour of iron in seawater, since Martin et al. [337–341] pointed out that thephytoplankton growth in oceanic water was limited by the deficiency of ironderived from the atmosphere, rather than the lack of nutrients in some oceanicregions, such as the equatorial Pacific, Gulf of Alaska, and Antarctic Ocean.This attractive hypothesis created a heated argument [291, 329–331, 342–345]and spurred the geochemical study of iron. For example, Zhuang et al. [346]reported recently that more than half of the iron in aeolian mineral dust existedin the form of iron (II), resulting in the enhancement of solubility of iron insurface water. In order to verify whether or not the iron deficiency contributesto the limitation of primary production and also to clarify the chemical speciesof iron, an accurate and rapid analytical method for determining iron seawateris essential.


Blain and Treguer [351] coupled a C18 column impregnated with ferrozine,a selective ligand for iron (II), to a spectrometer for online determination ofiron (II) in seawater in amounts down to 0.1 mM.

Shriadah and Ohzeki [350] determined iron in seawater by densitometryafter enrichment as a bathophenanthroline disulfonate complex on a thin layerof anion exchange resin. Seawater samples (50 ml) containing iron (II) andiron (III) were diluted to 150 ml with water, followed by sequential addition of20% hydrochloric acid (100 µl), 10% hydroxyl-ammonium chloride (2 ml), 5 Mammonium solution (to pH 3.0 for iron (III) reduction), bathophenanthrolinedisulfonate solution (1.0 ml), and 10% sodium acetate solution (2.0 ml) to givea mixture with a final pH of 4.5. A macroreticular anion exchange resin in the


chloride form, Amberlyst A27, was added, the resultant coloured thin layerwas scanned by a densitometer, and the absorbance measured at 550 nm.


A conventional analytical method, like solvent extraction–graphite furnaceatomic absorption spectrometric detection, requires a contamination-freetechnique. Moreover, it is time-consuming and troublesome, as litres of thesample solution must be treated because the dissolved concentration of ironin oceanic waters is extremely low (1 nmol/l = 56 ng/l). Martin et al. [341]recently found that the dissolved concentration of iron was less.

Atomic absorption spectrometry coupled with solvent extraction of ironcomplexes has been used to determine down to 0.5 µg/l iron in seawater [354,355]. Hiire [354] extracted iron as its 8-hydroxyquinoline complex. The sampleis buffered topH3–6andextractedwith a0.1%methyl isobutyl ketone solutionof 8-hydroxyquinoline. The extraction is aspirated into an air-acetylene flameand evaluated at 248.3 nm.

Moore [355] used the solvent extraction procedure of Danielson et al. [119]to determine iron in frozen seawater. To a 200 ml aliquot of sample was added1 ml of a solution containing sodium diethyldithiocarbamate (1% w/v) andammonium pyrrolidine dithiocarbamate (1% w/v) at pH to 4. The solution wasextracted three times with 5 ml volumes of 1,1,2 trichloro-1,2,2 trifluoroethane,and the organic phase evaporated to dryness in a silica vial and treated with0.1 ml Ultrex hydrogen peroxide (30%) to initiate the decomposition of organicmatter present. After an hour or more, 0.5 ml 0.1 M hydrochloric acid wasadded and the solution irradiated with a 1000 W Hanovia medium pressuremercury vapour discharge tube at a distance of 4 cm for 18 minutes. The iron inthe concentrate was then compared with standards in 0.1 M hydrochloric acidusing a Perkin-Elmer Model 403 Spectrophotometer fitted with a Perkin-Elmergraphite furnace (HGA 2200).

The coefficient of variation of analyses was 21% for seven subsamples con-taining 1.6 nmol Fe per litre, and 30% for eight subsamples at 0.6 nmol Fe perlitre. The detection limit was estimated to be 0.2 nmol Fe per litre.

The efficiency of the extraction procedure was tested using seawater spikedwith iron-59, which gave a recovery of 97%, and with stable iron, which gavea recovery of 86%.

5.28.3Chemiluminescence

The classical chemiluminescence method using a luminol–hydrogen peroxidesystem [341, 348] is thought to be a promising method for the shipboardanalysis of iron because it is highly sensitive to iron and requires only a small

5.29 Lanthanum 167

size detection device. However, iron (III) must be separated from the otherheavy metal ions, such as manganese (II), chromium (III), cobalt (II), andcopper (II) prior to detection, since the method is not specific to iron (III).To overcome these problems, Obata et al. [349] have developed an automatedanalytical method for determining iron in seawater using a closed flow systemwith a combination of a chelating resin concentration and chemiluminescencedetection.

This method is based on a combination of selective column extraction us-ing chelating resin, and improved chemiluminescence detection in a closedflow-through system. In this method, iron (III) in an acidified sample solutionis selectively collected on 8-quinolinol immobilised chelating resin and theneluted with dilute hydrochloric acid. The resulting eluent is mixed with luminolsolution, aqueous ammonia, and hydrogen peroxide solution successively, andthen the mixture is introduced into the luminescence cell. The iron concen-tration is obtained from the chemiluminescence intensity. The detection limitof iron (III) is 0.05 nmol/l when using an 18 ml seawater sample. This methodwas applied to ordinary oceanic waters and hydrothermal waters collected inthe North and South Pacific Oceans. O’Sullivan et al. [356] determined downto 0.1 nmol/kg ferrous iron in seawater by oxidation with oxygen, followed bydetermination by a catalysed Luminol chemiluminescence method.

Eirod et al. [352] determined sub-nanomolar levels of iron (II) and totaldissolved iron in seawater by flow injection analysis with chemiluminescentdetection in amounts down to 0.45 nmol/l.

5.28.4Voltammetry

Brendel and Luther [356] have described a solid-state voltammetric gold anal-yser microelectrode for the measurement of iron and magnesium in porewaters.

5.28.5Radioisotope Dilution

Radioisotope dilution using the chelating agent bathophenanthroline has beenused to determine down to 5 µg/l iron in seawater [357].

See Sects. 5.74.4–5.74.5 and 5.74.8–5.74.17.

5.29Lanthanum

See Sect. 5.74.16.


5.30Lead

5.30.1Atomic Fluorescence Spectroscopy

Bolshov et al. [358] used this technique to determine low lead concentrations.A detection limit of 0.05 pg ml–1 was achieved in studies with aqueous solutionsas reference using a graphite atomiser.

Laser-excited atomic fluorescence spectrometry has been used to determinedown to 1 ng/l of lead in seawater [359].


Ke and Lin [360] coupled a flame laser-enhanced ionisation technique withthe flow-injection analysis system to measure the traces of lead in aqueoussolution and in seawater. The flow-injection manifold is incorporated witha microcolumn packed with a C18 bonded silica. The chelating agent DDPAis used to form the Pb–DDPA complex, which may be sorbed in the micro-column and then eluted with methanol. The preconcentration lead is thendetected by the flame laser-enhanced ionization technique with either single-step or two-step extraction. At 5 and 15 ml volume-fixed sample loading, thedetection limits of 0.011 and 0.0033 ng/ml (11 and 3.3 ppt) and enrichmentfactors of 16 and 48 are achieved, respectively, using a two-step flow-injectionflame laser-enhanced ionisation. The sensitivity of the current system provesto be better by at least an order of magnitude than that of conventional flamelaser-enhanced ionisation method. The flow injection flame laser-enhancedionisation instrument also increases the tolerance of matrix interference. Theflame laser-enhanced ionisation signal is slightly reduced to 80% intensity as10 000 µg/ml (ppm) sodium and potassium matrixes are mixed in the leadsolution. The resistance to the alkali matrixes is approximately four times thefigure reported previously using a similar water-immersed probe as a flamelaser-enhanced collector. Finally, the flow injection flame laser-enhanced ion-isation technique was applied to detect the lead content in seawater, achievinga result of 0.0112 ± 0.00006 ng/mL (ppb), consistent with the certified value of0.013 ± 0.0005 ng/mL ∼ (ppb).


Various workers [361–365] have applied graphite furnace atomic absorptionspectrometry to the determination of lead in seawater.

5.30 Lead 169

The large amount of sodium chloride in seawater samples causes nonspecificabsorption [366–370], which can only be partially compensated by backgroundcorrection. In addition the seawater matrix may give rise to chemical as wellas physical interferences related to the complex physico-chemical phenom-ena [371–373] associated with vaporization of metals and of the matrix itself.

Severalmatrixmodifiers,whichalter thedryingorcharringpropertiesof thesample matrix, have been tested [374–378] to reduce nonspecific absorption.However, the matrix modification methods do not permit determinations ofthe indigenous lead in seawater because of the relatively high detection limitand poor precision, and yet gross chemical manipulations of the samplesshould be avoided to prevent contaminations which can be dramatic when theanalyte is present at µg/l or sub-µg/l level.

With the temperature-controlled heating method described by Lundgrenet al. [397], the heating rate can be chosen independently of final temperature,thus permitting a selective volatilisation. However, this method cannot be usedsuccessfully for the determination of lead in strong sodium chloride solutionslike seawater, because the temperature at which atomisation of lead is rapidcoincides with the volatilization temperature of sodium chloride. Ashing ofseawater samples by a hydrogen diffusion flame [380] was successful in thedirect determination of iron, nickel, and copper but cannot be applied to leadbecause hydrogen is not sufficiently effective as a suppressor for the lead–NaClsystem.

In order to overcome the problem of the high nonspecific absorption, al-ternative procedures have been tested, which involve prior separation of thetrace metals from the salt matrix. Examples of extraction of trace metals fromseawater as chelates with subsequent determination by electrothermal atomicabsorption spectrometric procedures have been described [381,382], but theseand similar methods are seldom effective and satisfactory when the matrix isvery complex and the analyte concentration very low.

In contrast, the coupling of electrochemical and spectroscopic techniques,e.g., electrodeposition of a metal followed by detection by atomic absorptionspectrometry, has received limited attention. Wire filaments, graphite rods,pyrolytic graphite tubes, and hanging drop mercury electrodes have beentested [383–394] for electrochemical preconcentration of the analyte to be de-termined by atomic absorption spectroscopy. However, these ex situ precon-centration methods are often characterised by unavoidable irreproducibility,contaminations arising from handling of the support, and detection limitsunsuitable for lead detection at sub-ppb levels.

These drawbacks could be certainly avoided by performing in situ deposi-tion. The sole attempt in this direction was made by Torsi [395], who set up anapparatus which permitted both in situ electrochemical preconcentration ofthe analyte from a flowing solution and almost complete suppression of matrixeffects because the matrix could be removed by suitable washing. The feasibil-ity of this approach was successfully tested with respect to lead determinations


in different salt solutions (mainly ammonium acetate), and some preliminaryresults were reported for real samples of seawater.

Torsi et al. [395] have carried out a systematic investigation to establish thepotential value of such an apparatus. The apparatus is basically an electrother-mal device in which the furnace (or the rod) is replaced by a small cruciblemade of glassy carbon. Figure 5.10 provides an overall view of the apparatus.Figure 5.11 shows a block diagram of the electrolysis circuit; the crucible (6)acts a cathode, while the anode is a platinum foil dipped into either the samplesolution reservoir (1) or the washing solution reservoir (2). Pre-elecrolysis wasperformed at constant current with a 500 V dc variable power supply (5). Underthese conditions, the cathode potential is not controlled, so that other metalscan be codeposited with lead.

A typical measurement was performed as follows. The feeder was loweredinto the crucible and the sample solution (seawater) was allowed to flow un-der an inert atmosphere with the suction on. A constant current was appliedfor a predetermined time. When the pre-electrolysis was over, the flow waschanged from the sample to the ammonium acetate washing solution, whilethe deposited metals were maintained under cathodic protection. Ammoniumacetatewas selected for its lowdecomposition temperature, anda0.2 ml l–1 con-centration was used to ensure sufficient conductivity. At this point the feedertip was raised to the highest position and the usual steps for an electrothermalatomic absorption spectrometry measurement were followed: drying for 30 s at90 ◦C,ashing for30 s at 700 ◦C,andatomization for8 sat 1700 ◦C,withmeasure-ment at 283.3 nm. The baseline increases smoothly with time as a consequenceof an upward lift of the crucible caused by thermal expansion of the material.

A deviation from linearity is observed in the calibration curve at higherlead concentrations. The estimated value for the original sample was found tobe 0.51 µg/l, with confidence limits at the 95% confidence level of ±0.036 µg/l,compared with a value of 0.65±0.08 µg/l obtained by anodic scanning voltam-metry. This value is well within the normal range reported in the literature forthenatural leadcontentofunpollutedseawater.Adetection limitof 0.03 ng ml–1

was obtained.Halliday et al. [396] have described a simple rapid graphite furnace method

for the determination of lead in amounts down to 1 µg/l in polluted seawa-ter. The filtered seawater is diluted with an equal volume of deionised water,ammonium nitrate added as a matrix modifier, and aliquots of the solutioninjected into a tantalum-coated graphite tube in an HGA-2200 furnace atom-iser. The method eliminates the interference normally attributable to the ionscommonly present in seawater. The results obtained on samples from the Firthof Forth (Scotland, UK) were in good agreement with values determined byanodic stripping voltammetry.

Hirao et al. [964] concentrated lead in seawater using a chloroform solu-tion of dithizone and determined it in amounts down to 40 µg/l by graphitefurnace atomic absorption spectrometry. Lead in 1 kg acidified seawater was

5.30 Lead 171

Figure 5.10. Overall view of the apparatus. 1 Vitreous carbon crucible. 2 Graphite rod.3 Water-cooled steel column electrical leads. 4 Plexiglas cover. 5 Feeder. 5a Feeder tip.6a–c Slide knobs. 7a,b Washing and sample solution reservoirs. The glassy carbon cru-cible (1) is 8 mm high, 5 mm OD, 3 mm ID, and 6 mm deep (Carbone Lorraine, Paris).The graphite rods (2) hold the crucible firmly in position. Water-cooled stainless steelcolumns (3) force the graphite rods against the crucible by means of two embedded screws,and also serve as electrical leads. The Plexiglas box (4) maintains the controlled inert at-mosphere required to avoid drastic reduction of the absorption signal due to oxygen. Thesolution feeder (5) can be moved up and down by means of knob (6a) into a metal blockattached to the upper part of the Plexiglas box. A three-way stopcock at the top of the cylin-der connects the feeder tip (5a) to the washing and sample solution reservoirs via Teflontubing (1.5 mm OD, 0.8 mm ID). Additional slide knobs (6b and c) move the feeder in thehorizontal plane, facilitating micrometric positioning of the feeder tip. Source: [395]

equilibrated with lead of a known radioactivity, extracted with dithizone inchloroform, back-extracted with 0.1 M hydrochloric acid and subjected tographite furnace atomic absorption spectrometry by a two-channel spectrom-eter. Recovery yield of lead was found to be 60–90% from the radioactivity


Figure 5.11. Electrolysis circuit layout. 1,2 Sample and washing solution compart-ments. 3 Three-way stopcock. 4 Ammeter.5 500 V Variable dc power supply. 6 Cru-cible. Source: [395]

of 212Pb in the back-extract. Lead concentrations were thus determined withabout 10% precision.

Ohta and Suzuki [397] investigated the electrothermal atomisation of leadfor accurate determination of lead in water samples. Thiourea served to lowerthe atomisation temperature of lead and to eliminate the interferences fromchloride matrix. The addition of thiourea also allowed the accurate determi-nation of lead irrespective of its chemical form. The absolute sensitivity (1%absorption) was 1.1 ×10–12 g of lead. The method permits the direct rapiddetermination of lead in water samples including seawater.

No severe interference was noted in this method for arsenic, bismuth, cal-cium, copper, iron, magnesium, antimony, selenium, tin, and tellurium.

Sturgeon et al. [398] applied in situ metal trapping to the determination oflead in seawater.

In this method, inorganic lead in seawater samples are converted to tetra-ethylead using sodium tetraethylboron (NaB(C2H5)4) which is then trappedin a graphite furnace at 400 ◦C. Quantitation is achieved by using a simplecalibration graph prepared from aqueous standards. An absolute detectionlimit of (3σ) 14 pg is achieved. Precision of determination at 100 pg ml–1 is 4%relative standard deviation.

Izumi et al. [948] used X-ray absorption fine structure combined withfluorescence spectrometry to monitor trace amounts of lead in seawaters.


Clem [399] has described an electrochemical cell in which rapid deoxygenationof the sample solution is achieved by allowing a jet of nitrogen to impinge on thesurface of the liquid, while the cell is rotated to maintain the solution as a thin

5.30 Lead 173

film on the cell wall; 15 ml of solution can be deoxygenated in 1 to 1.5 minutes.Stirring during analysis is by periodic reversal of the cell rotation. The cellhas been used for determining, by anodic-stripping voltammetry, 11.2 and 4.1parts per 109 of lead in acidified seawater (pH 2), and for the amperometrictitration (with lead) of organic ligands in non-acidified seawater.

Early work [400] on the application of cyclic and anodic stripping voltam-metry to the determination of lead showed a poor correlation between peakcurrent values and Pb11 concentration at high pH values. This is due to the lowelectrochemical activity of PbOH.

Acebal et al. [401] discussed the quantitative behaviour of lead (and copper)when voltammetric determinations are done at mercury film electrodes andhanging mercury drop electrodes. The samples were collected in polyethylenebottles and, generally, were not acidified immediately after collection. Thismight place some doubt on the results reported.

Voltammetric measurements were done with a PAR Electrochemical System(Model 174-A) and a saturated calomel reference electrode, a platinum wireauxiliary electrode and a glassy carbon rod (PAR 0333) coated with a mercuryfilm as the working electrode. A Pyrex glass cell was used for measurementswith the hanging drop mercury electrode; either this cell or a Teflon cell wasused with the mercury film electrode. No advantage concerning adsorption orcontamination was found when a Teflon cell was used for seawater at pH 2.Stirring was done magnetically. Nitrogen with a maximum oxygen content of10 ppm was used as purging gas. Mercury (II) nitrate used to form in situ filmson theglassy carbonrodwasprepared fromtridistilledmercuryandnitric acid.

This work showed that application of the linear sweep mode at pH1.5 is a fastand reliable way of dealing with interferences caused by organic materials inpolluted water. The lower sensitivity of this mode limits its use to lead contentsexceeding 0.1 µg/kg but such levels are commonly reached in polluted waters.

The determination of lead in seawater collected from 14 stations in Guan-abara Bay gave values between 0.07 and 5.5 µg/kg.

Quentel et al. [402] studied the influence of dissolved organic matter in thedetermination of lead in seawater by anodic stripping voltammetry.

Svensmark [403] gives details of equipment and procedure for the rapiddetermination of lead by a modification of anodic stripping voltammetry, usingstaircase voltammetry at high scan rates to strip the lead plated on a rotatingmercury-film electrode. This allowed rapid determination of lead without theneed for de-oxygenation, rest periods, electrode rotation or stirring. Lead ata concentration as low as 0.1 µg/l could be determined in less than 4 minutes.Results obtained on a sample of seawater are presented.

To determine down to 6 ppt of lead in seawater Wu and Batley [404] used ab-sorptive stripping voltammetry with ligand competition using xylenol orange.

Garcia-MoncoCarraet al. [405]havediscussed theuseofa“hybrid”mercuryfilm electrode for the voltammetric analysis of lead (and copper) in acidifiedseawater.


5.30.5Mass Spectrometry

Flegal and Stukas [406] described the special sampling and processing tech-niques necessary for the prevention of lead contamination of seawater samples,prior to stable lead isotopic ratio measurements by thermal ionisation massspectrometry. Techniques are also required to compensate for the absence ofan internal standard and the presence of refractory organic compounds. Theprecision of the analyses is 0.1–0.4% and a detection limit of 0.02 ng/kg allowsthe tracing of lead inputs and biogeochemical cycles.

5.30.6Miscellaneous

Baena et al. [949] studied the speciation of lead in environmental waters.Ultraviolet spectroscopy has been applied to the determination of lead and

lead speciation studies [407]. Scaule and Patterson [408] used isotope dilution-mass spectrometry to determine the lead profile in the open North PacificOcean.

See also Sects. 5.74.4–5.74.16.

5.31Lithium


Benzwi [409] determined lithium in Dead Sea water using atomic absorptionspectrometry. The sample was passed through a 0.45 µm filter and lithium wasthen determined by the method of standard additions. Solutions of lithium inhexanol and 2-ethylhexanol gave greatly enhanced sensitivity.

5.31.2Gel Permeation Chromatography

Rona and Schmuckler [410] used gel permeation chromatography to separatelithium from Dead Sea brine. The elements emerged from the column in theorder potassium, sodium, lithium, magnesium, and calcium and it was possibleto separate a lithium-rich fraction also containing some potassium and sodiumbut no calcium and magnesium.


Wiernik and Amiel [411] used neutron activation analysis to measure lithiumand its isotopic composition in Dead Sea brines.


5.32 Lutetium 175

5.32Lutetium

See Sect. 5.49.14.

5.33Magnesium

5.33.1Gravimetric Method

Das et al. [412] carried out a direct gravimetric determination of magnesiumin seawater by precipation with N-benzoyl-N-phenylhydroxylamine. The pre-cipitate is weighed as Mg(C13)H10O2N)2. The coefficient of variation for 5 mgof magnesium was 0.55%. Magnesium could be determined in the presenceof barium or strontium; coprecipitation of calcium was avoided by addingtartrate, nickel, cobalt, copper, mercury, and zinc were masked with cyanideand tartrate. Phosphate, oxalate, fluoride, and EDTA interfered. Tin, iron, alu-minium, and beryllium were separated by prior precipitation with N-benzoyl-N-phenylhydroxylamine at pH 0.5 to 1.0, 3.5, 4.0 and 5.5, respectively.


Atomic absorption spectrometry has been used to determine magnesium inseawater [413–415].

See also Sects. 5.74.1 and 5.74.2.

5.34Manganese

The natural water chemistry of manganese, which is an important trace el-ement both biologically and geologically, is complicated by non-equilibriumbehaviour. From thermodynamic considerations manganese dioxide (man-ganese (IV)) is expected to be the stable form of manganese in seawater [416].However, seawater contains a significant quantity of dissolved manganese (II)which is only slowly oxidised (Murray and Brewer [417]). Investigations of theoxidation of manganese (II) have shown that the process is autocatalytic, theproduct being a solid manganese oxide phase whose composition depends onthe reaction conditions [417–420]. The heterogenous nature of the oxidationprocess can thus account for the extremely slow oxidation of manganese (II)in seawater where concentrations of particulate matter can be relatively small.

Estuaries, in contrast, appear to be important sites for manganese redox re-actions.Manganesemaximahavebeenobserved in several estuaries [421–423]


and it has been suggested that these maxima result from a recycling of pre-cipitated manganese [423]. The proposed mechanism is essentially a redoxcycle in which dissolved manganese (II) is oxidised into the water column andprecipitated. Reduction in anoxic sediments results in the subsequent releaseof manganese (II) to the water column.

The details of estuarine manganese chemistry are far from clear, howeverSholkovitz [424] notes that while adsorption onto colloidal humic acids orhydrous iron oxides is a major factor controlling the removal of many tracemetals from estuarine waters, manganese does not conform to this patternas it is known to behave independently of iron and associates only weaklywith organic matter. Detailed investigation of estuarine manganese reactionsrequires analytical methods specific to the species involved; a requirement metonly by electrochemical methods at natural concentration levels. Davison [425,426] has described the use of direct polarographic methods in the analysis ofmanganese in lake waters in the concentration range 0.1–5 mg/l.

Manganese is of particular interest becauseof its central role inmanymarinegeochemical processes and involvement in biological systems.

Manganese and many other trace metals are present in open ocean waters atconcentrations in the order of nmol/l or less, and it is only relatively recently,when adequate contamination control measures have been applied duringsampling and measurements, that accurate data have been obtained.

Manganese is a geochemically active element in the ocean. The dissolvedmanganese is easily precipitated by oxidation to manganese (IV) oxide, whichacts as a powerful scavenger for trace elements. The solidified manganesein sediments is reduced to manganese (II) and is regenerated into the watercolumn under mild reducing conditions, for example in the oxygen minimumzone and the near-shore anoxic sediments. In recent years, it has also beenfound that a copious amount of manganese is injected into the deep waters byhydrothermal emanations through the active ocean crusts. Therefore it is veryimportant to clarify the distribution of manganese in seawater to understandmarine geochemistry. Furthermore, manganese is thought to be a promisingelement as a chemical tracer for probing the hydrothermal activities if it canbe analysed easily and quickly onboard ship.


Olafsson [427] has described a semiautomated determination of manganese inseawater using leucomalachite green. The autoanalyser had a 620 nm interfer-ence filter and 50 minute flow cells. Findings indicated initial poor precisionwas related to pH, temperature, and time variations. With strict controls onsample acidity and reaction conditions, the semiautomated method had highprecision, at least as good as that achieved by preconcentration and atomicabsorption procedures, and provided precise, rapid, shipboard information

5.34 Manganese 177

on the continental distribution of manganese and anomalies associated withgeothermal sea-floor activity. The method was not suitable for estuarine sam-ples, nor quite sensitive enough for study of the open ocean manganese distri-bution.

Brewer and Spencer [428] have described a method for the determinationof manganese in anoxic seawaters based on the formulation of a chromophorwith formaldoxine to produce a complex with an adsorption maximum at450 nm. Sulfide (50 µg/l), iron, phosphate (8 µg/l), and silicate (100 µg/l) donot interfere in this procedure. The detection limit is 10 µg/l manganese.


Biddle and Wehry [429] carried out fluorometric determination of man-ganese II in seawater via catalysed enzymic oxidation of 2,3-diketogluconate.The detection limit was 8 µmol l–1 Mn (II).


Graphite-furnace atomic absorption spectrometry, although element-selectiveand highly sensitive, is currently unable to directly determine manganese atthe lower end of their reported concentration ranges in open ocean waters.Techniques that have been successfully employed in recent environmentalinvestigations have thus used a preliminary step to concentrate the analyte andseparate it from the salt matrix prior to determination by atomic absorptionspectrometry.

The determination of manganese in seawater using graphite furnace atomicabsorption spectrometry has been investigated by many workers [129, 430–437]. If the seawater matrix is atomized along with the analyte, the result isa large background signal which is often beyond the correcting capabilities ofcurrent instrumentation. The presence of large amounts of chlorides has alsobeen shown to provide interferences [438,439], usually making direct analysisdifficult.

To ameliorate such problems, most workers either have used matrix mod-ification or have extracted the metal from the seawater matrix [129, 430,432–434, 437]. Few workers have been successful with the direct determi-nation in seawater after volatilisation of the matrix during the char programstep [430, 431, 435, 436, 439, 440]. Slavin and Manning [441–443] have shownthat by using a furnace at steady-state temperature (the L’vov platform), the in-terference of chloride on manganese determination was greatly reduced as longas the background signal was within limits that the deuterium arc backgroundcorrector could handle.


Segar and Gonzalez [431] attributed the reduced sensitivity to manganesein seawater matrix to covolatilisation of some manganese with the salt matrix.More recent work suggests that this reduced sensitivity is due to vapour-phasebinding of a portion of the manganese by chloride.

Segar and Cantillo [436] developed a direct method for manganese in sea-water with a detection limit for manganese of about 0.3 µg/l. Only ordinarygraphite tubes were available, and they found that as the tube aged, the ana-lytical signal fell linearly at a rate of 50% per 100 firings. Since variations insalinity produced relatively large changes in signal, the method of standardadditions was required.

McArthur [433] preferred to use ammonium nitrate matrix modification todetermine manganese in seawater. Most of his paper discussed the charringprocess. He found considerable salinity dependence if charring was too rapid.There was considerable change in the salinity dependence with the age of thetubes.

Kingston et al. [129] resorted to extraction on Chelex 100 followed by strip-ping with nitric acid. The ammonium nitrate matrix modification techniquewas used by Montgomery and Peterson [437] for the determination of man-ganese in seawater. They showed that the pyrolytically coated tubes they useddeteriorated very rapidly when they used the combination of ammonium ni-trate and seawater. Manganese was determined in seawater (with copper andcobalt) by Hydes [465] after adding 1% ascorbic acid to the sample. He usedthe Perkin-Elmer HGA-2100 furnace and found significant loss of manganesefrom seawater between 600 ◦C and 900 ◦C.

The direct furnace method of Sturgeon et al. [430] for manganese was verysimilar to the method of Segar and Cantillo [472]. The Sturgeon detectionlimit was 0.22 µg/l for manganese in seawater, using 20 µl samples in the HGA-2200 furnace and pyrolytically coated graphite tubes. They found a loss insensitivity during the life of the tubes. They had to use the method of additionsto accommodate small residual interference.

Procedures using chelation followed by extraction have been describedfor manganese using the 8-hydroxy-quinoline-chloroform system [432, 444].Dithiocarbamate systems can simultaneously extract manganese, along withother trace metals under suitable conditions [445–447].

Statham [448] has optimised a procedure based on chelation with am-monium dithiocarbamate and diethylammonium diethyldithiocarbamate forthe preconcentration and separation of dissolved manganese from seawaterprior to determination by graphite furnace atomic absorption spectrometry.Freon TF was chosen as solvent because it appears to be much less toxic thanother commonly used chlorinated solvents, it is virtually odourless, has a verylow solubility in seawater, gives a rapid and complete phase separation, and isreadily purified. The concentrations of analyte in the back-extracts are deter-mined by graphite furnace atomic absorption spectrometry. This procedureconcentrates the trace metals in the seawater by a factor of 67.3.

5.34 Manganese 179

When a 350 ml seawater sample was spiked with 54Mn and taken through thechelation, extraction, and back-extraction procedures, the observed recoveryof the radio-tracer was 100.6%. Estimates of detection limits for manganesebased on sets of both shipboard and shore laboratory separations are of theorder of 0.1 nmol/l. The accuracy of the technique is demonstrated by data fromthe ICES fifth-round intercalibration exercise for trace metals in seawater [449].

Klinkhammer [432] has described a method for determining manganese ina seawater matrix at concentrations ranging from about 30 to 5500 ng/l. Thesamples are extracted with 4 nmol/l 8-hydroxyquinoline in chloroform, andthe manganese in the organic phase is then back-extracted into 3 M nitric acid.The manganese concentrations are determined by graphite furnace atomicabsorption spectrophotometry. The blank of the method is about 3.0 ng/l, andthe precision from duplicate analyses is ±9% (1 SD).

The theoretical yield of the method is less than 100%, as only 80–90% of theaqueous phase is removed after back-extraction. The actual yield obtained by54Mn counting was 69.5±7.8%, and this can be allowed for in the calculation ofresults. Environmental Protection Agency standard seawater samples of knownmanganese content (4370 ng/l) gave good manganese recoveries (4260 ng/l).

Bender and coworkers [450, 451] determined total and soluble manganesein seawater. The samples were collected into 500 ml polyethylene bottles. Allsamples were brought to pH 2 with nitric acid free of trace metals, and storedin individual zip-lock plastic bags to minimise contamination.

When the samples were returned to the laboratory the pH was adjustedto approximately pH 8 using concentrated ammonia (Ultrapure, G. Freder-ick Smith). Chelating cation exchange resin in the ammonia form (20 ml;Chelex 100, 100–200 mesh, Bio-Rad) was added to the samples and they werebatch extracted on a shaker table for 36 hours. The resin was decanted intocolumns, and the manganese eluted using 2 N nitric acid [129]. The eluantwas then analysed by graphite furnace atomic absorption spectrophotometry.Replicate analyses of samples indicate a precision of about 5%.

Graphite furnace atomic absorption spectrometry with the L’vov platformand Zeeman background correction has been applied to the determination ofdown to 0.02 µg/l manganese in seawater [452].

Lan and Alfassi [453] determined manganese in seawater in amounts downto 50 ppt using 50 µl sample by graphite furnace atomic absorption spectrom-etry.

To determine manganese [452, 454], several factors had to be controlledcarefully to obtain reliable results against simple standards that were indepen-dent of salinity and variations in matrix composition. Use of the L’vov platformand integration of the absorbance signal reduced the sensitivity to matrix com-position. Pyrolytically coated graphite reduced variations that depend uponthe life of the tubes. The tubes appeared to fail by intercalation of the sodiumor sodium chloride matrix. The char temperature must not vary outside therange 1100–1300 ◦C. Zeeman background correction permitted use of larger


seawater samples. The detection limit of the procedure using 20 µl samples was0.1 µg/l (2 pg) manganese. By use of the Zeeman background corrector, lessthan 0.02 µg/l manganese was detected in seawater using a 75 µl sample.

5.34.4Polarography

Knox and Turner [454] have described a polarographic method for mangane-se (II) in estuarine and seawaters which covers the lower concentration range10–300 µg/l. The method, which is specific to manganese (II) and its labilecomplexes, is used in conjunction with a colorimetric technique to comparethe levels of manganese (II) and total dissolved manganese in an estuarinesystem. They showed that polarographically determined manganese (II) canvary widely from 100% to less than 10% of the total dissolved manganese, de-termined spectrophotometrically at 450 nm by the formaldoxine method [455]calibrated in saline medium to overcome any salt effects. It is suggested that themanganese not measured by the polarographic method is present in a colloidalform.


Neutron activation analysis has been used to determine total manganese inseawater.

Wiggins et al. [456] used neutrons from the thermal column of a 10 kWpool-type research reactor and from a 120 µg Cf source to study the promptphoton emission resulting from neutron capture in magnesium nodules (ter-romanganese oxides) from the ocean floor. Spectra were recorded with a Ce(Li)detector and a 1024-channel analyser. Complex spectra were obtained by ir-radiation of seawater, but it was possible to detect and estimate manganesein nodules in a simulated marine environment by means of the peaks at 7.00,6.55, 6.22, and 6.04 µV.

See also Sects. 5.74.4–5.74.6, 5.74.8–5.74.12, and 5.74.14–5.74.17.

5.35Mercury


Hedeishi and McLaughlin [457] have reported the application of the Zeemaneffect for the determination of mercury. Atomic absorption and atomic fluo-rescence techniques using closed system reduction–aeration have been appliedwidely to determine mercury concentrations in natural samples [458–472].

5.35 Mercury 181

Typical of these methods is that of Topping and Pirie [468], in which themercury is concentrated by drawing air for 5 h (600 ml/min) through a mixtureof the sample (4 litres) and 20% stannous chloride solution in 5 M hydrochlo-ric acid (45 ml), and absorbing mercury vapour from the air stream in 20 ml2% potassium permanganate solution: 50% (v/v) sulfuric acid (1:1). To theabsorption solution was added 15 ml of the stannous chloride solution, andthis mixture was aerated at 2 l/min. The air and mercury vapour are passedthrough a 15 cm gas cell (with silica windows) in an atomic absorption spec-trophotometer for measurement at 253.65 nm. Samples containing down to2 ng/l of Hg could be analysed by this procedure.

Olafsson [472] described a similar procedure, in which the sample (450 ml)is acidified with nitric acid, aqueous stannous chloride is added, and themercury is entrained by a stream of argon into a silica tube wound externallywith resistance wire and containing pieces of gold foil, on which the mercury isretained. The tube and its contents are then heated electrically to about 320 ◦Cand the vaporised mercury is swept by argon into a 10 cm silica absorptioncell in an atomic absorption spectrophotometer equipped with a recorder. Theabsorption (measured at 253.7 nm) is directly proportional to the amount ofmercury in the range 0–24 ng per sample.

In many applications, such as the analysis of mercury in open ocean seawa-ter, where the mercury concentrations can be as small as 10 ng/l [468,472–476],a preconcentration stage is generally necessary. A preliminary concentra-tion step may separate mercury from interfering substances, and the low-ered detection limits attained are most desirable when sample quantity islimited. Concentration of mercury prior to measurement has been com-monly achieved either by amalgamation on a noble-metal metal [460, 467,469, 472], or by dithizone extraction [462, 472, 475] or extraction with sodiumdiethyldithiocarbamate [475]. Preconcentration and separation of mercuryhas also been accomplished using a cold trap at the temperature of liquidnitrogen.

Fitzgerald et al. [477] have described a method based on cold-trap precon-centration prior to gas-phase atomic absorption spectrometry for the deter-mination of mercury down to 2 ng/l in seawater.

The cold trap is created by immersing a glass U-tube packed with glass beads(80/100 mesh) in liquid nitrogen. After reduction, purging, and trapping, themercury is removed from the glass column by controlled heating, and thegas-phase absorption of eluting mercury is measured. This procedure hasbeen employed for both shipboard and laboratory analyses of mercury inseawater.

The mercury analyses were conducted using a Coleman Instruments mer-curyanalyser (MAS-50)equippedwitharecorder.Theaqueous sample solutionwas contained in a 250 ml Pyrex glass bubbler placed at one end of a samplingtrain employing nitrogen as the purging and carrier gas. A schematic diagramof the entire system is shown in Fig. 5.12.


Figure 5.12. Hg cold-trap preconcentration system for the determination of mercury bygas-phase absorption. Source: [477]

Fitzgerald et al. [477] showed that the most significant quantities of mercuryoccurred in the waters of the Atlantic Ocean’s continental shelf and slope (21–78 ng/l), compared with open ocean samples (2–11 ng/l).

These workers distinguished between inorganic mercury obtained by di-rect analysis on the sample as received, and organic mercury (the differencebetween total mercury obtained upon ultraviolet irradiation of the sample andinorganic mercury).

Olafsson [478] has reported on the results obtained in an internationalintercalibration for mercury in seawater. Sixteen countries participated in thisexercise, which involved analysis of a seawater and seawater spiked with 15.4and 143 ng/l mercury. The results show good accuracy and precision in therecovery of mercury for the majority of calibrations, but serious errors in thelow-level determinations on the seawater.

Since the intercalibration samples had been ultraviolet-irradiated, the ma-jority of participants in this exercise preferred to analyse the samples withoutany pretreatment. Ten did, however, employ oxidising pretreatment, and withthree exceptions the results suggest that this approach should be taken withgreat caution.

Half of the participants preconcentrated mercury from the seawater priorto determination, eleven by amalgamation on gold, one by amalgamation onsilver, two by collection into oxidising solutions, one by organic extraction,and one by ion exchange chromatography.

Reduction to metallic mercury was used by an overwhelming propor-tion of the participants, with stannous chloride as reductant in all but onecase in which sodium borohydride was used. In all cases but four, the par-ticipants used cold-vapour atomic absorption for final determination. Thismakes comparison of detection techniques difficult, but the good results ob-

5.35 Mercury 183

tained by cold-vapour atomic fluorescence are of interest and the spuriousresults obtained by differential pulse anodic stripping voltammetry may beindicative of the risk of mercury contamination in polarographic laborato-ries.

Filippelli [479] determined mercury at the subnanogram level in seawaterusing graphite furnace atomic absorption spectrometry. Mercury (II) was con-centrated using the ammonium tetramethylenedithiocarbamate (ammoniumpyrrolidine-dithiocarbamate, APDC)-chloroform system, and the chloroformextract was introduced into the graphite tube. A linear calibration graph wasobtained for 5–1500 ng of mercury in 2.5 ml chloroform extract. Because ofthe high stability of the HgII-APDC complexes, the extract may be evapo-rated to obtain a crystalline powder to be dissolved with a few microlitres ofchloroform.

About 84% of mercury was recovered in a single extract (97% in two extrac-tions). The calibration graph was prepared by plotting the peak height againstamount of mercury added to 500 ml distilled water. The optimized experi-mental conditions are as follows: lamp current, 6 mA; wavelength, 253.63 nm;drying, 100 ◦C for 10 s; ashing, 200 ◦C for 10 s; atomisation, 2000 ◦C for 3 s; andpurge gas, nitrogen “stopped flow”.

Thecoefficient variationof thismethodwasabout 2.6%at the1 µg/lmercurylevel. The calibration graph is linear over the range 5–1500 µg mercury.

Blake [480] has described a method for determination of trace amounts ofmercury, with a limit of detection of less than 2 ng/l in fresh and saline waters.It was based on generating mercury vapour from the sample by reduction,together with trapping on gold mesh, subsequent desorption and measurementby cold vapour atomic absorption spectrometry. Checks on the precision andrecovery of the method with respect to inorganic mercury are described, andthe recovery of methyl mercury was also investigated. The performance of themethod was within the limits implied in the requirements of the HarmonisedMonitoring scheme and the application of EC Directives concerned with waterquality monitoring.

Gill and Fitzgerald [481] determined picomolar quantities of mercury inseawater using stannous chloride reduction and two-stage amalgamation withgas-phase detection. The gas flow system used two gold-coated bead columns(the collection and the analytical columns) to transfer mercury into the gascell of an atomic absorption spectrometer. By careful control and estimationof the blank, a detection limit of 0.21 pM was achieved using 2 l of seawater.The accuracy and precision of this method were checked by comparison withaqueous laboratory and National Bureau of Standards (NBS) reference mate-rials spiked into acidified natural water samples at picomolar levels. Furtherstudies showed that at least 88% of mercury in open ocean and coastal seawaterconsisted of labile species which could be reduced by stannous chloride underacidic conditions.



Bloxam et al. [482] used liquid chromatography with an inductively coupledplasma mass spectrometric detector in speciation studies on ppt levels ofmercury in seawater.

Turyan and Mandler [483] determined ppt levels of mercury in seawater byfirst converting mercury salts to elemental mercury using stannous chloride,the mercury was then trapped on gold deposited on platinum gauze andreleased by heating prior to determination by inductively coupled plasmamass spectrometry.

Debrak and Denoyer [484] determined mercury at the ppt level in seawaterby the addition of tin chloride to produce hydrogen vapour, and trapping ongold-platinum gauze, prior to heating and detection of the mercury releasedby inductively coupled plasma mass spectrometry.


Watling [491] has described an analytical technique for the accurate determi-nation of mercury at picogram per litre levels in fresh and seawater. Mercury,released by tin (II) chloride reduction of water samples is amalgamated ontosilver wool contained in quartz amalgamation tubes. The wool is then heatedand the mercury thus released is flushed by argon into a plasma where itis excited. The emission signal thus produced results in a detection limit of3 ×10–17 g and an analytical range 1 ×10–14 g–1 ×10–7 g.

5.35.4Atomic Emission Spectrometry

Wrembel [485] gives details of a procedure for the determination of mercury inseawater by low-pressure ring-discharge atomic emission spectrometry withelectrolytic preconcentration on copper and platinum mesh electrodes. Be-tween 40 ± 5 (open sea) and 50 ± 8 (shore area) µg/l mercury was found inBaltic sea waters.

Wrembel and Pajak [486] evaporated mercury from natural water sampleswith argon and amalgamated the mercury with a gold foil. The mercury wasexcited in a ring-discharge plasma and determined by atomic emission spec-troscopy. The method was applied to the determination of mercury in seawaterin the range 0.01–1.0 µg/l.

5.35.5Colloid Flotation

Voyce and Zeitlin [487] have used adsorption colloid flotation to determinemercury in seawater. The sample 500 ml is treated with concentrated hy-

5.35 Mercury 185

drochloric acid; an aqueous solution of cadmium sulfate and a fresh aque-ous solution of sodium sulfate are added. The pH is adjusted to pH 1.0 andthen poured into a flotation cell with a nitrogen flow of 10 ml/min. Ethanolicoctadecyltrimethylammonium chloride is injected and the froth dissolved inaqua regia in a flameless atomic absorption cell.

Following reduction of mercury with stannous chloride the mercury vapouris flushed from the system.

To determine organically bound mercury, the sample is treated (500 ml)with 0.5 M sulfuric acid aqueous potassium permanganate and set aside for24 h. Aqueous hydroxylammonium chloride is added and the determinationcompleted as above. The amount of mercury in the samples is calculated byreference to the standard absorptions. Average recoveries of 0.05 µg mercurywere 88%.

Agemian and Da Silva [488] have described an automated method for totalmercury in saline waters.

Bioassay methods have been used to obtain estimates of low mercury con-centrations (5–20 µg/l) in seawater [489]. This method is useful for detectingcomparatively small enhancements over background mercury concentrationsin estuarine and seawater.

This method consists of suspending for a standard time 70 mussels (Mytilusedulis), each of a standard weight, in a plastic coated wire cage 2 m belowthe surface. Mercury in the mussels was determined by cold vapour atomic

Figure 5.13. Relationship between mean total mercury concentration in water at the cagepositions (A–E) and the mercury loadings per mussel after various exposure times. © =28 days; � = 55 days;� = 106 days; × = 153 days. Source: [489]


absorption spectrometry [468, 490]. The procedure is calibrated by plottingdetermined mercury content of mussels against the mercury content of theseawater in the same area (Fig. 5.13).

5.35.6Miscellaneous

Other techniques that have been used include subtractive differential pulsevoltammetry at twingold electrodes [492], anodic strippingvoltammetryusingglassy-carbon electrodes [495, 496], X-ray fluorescence analysis [493], andneutron activation analysis [494].


5.36Molybdenum


Kawabuchi and Kuroda have concentrated molybdenum by anion exchangefrom seawater containing acid and thiocyanate [497] or hydrogen perox-ide [497,498], and determined it spectrophotometrically. Korkisch et al. [499]concentrated molybdenum from natural waters on Dowex 1-X8 in the presenceof thiocyanate and ascorbic acid.

Kuroda and Tarui [498] developed a spectrophotometric method for molyb-denum based on the fact that MoVI catalyses the reduction of ferric iron bydivalent tin ions. The plot of initial reaction rate constant versus molybdenumconcentration is rectilinear in the range 0.01–0.3 mg/l molybdenum. Severalelements interfere, namely, titanium, rhenium, palladium, platinum, gold, ar-senic, selenium, and tellurium.

In a method described by Kiriyama and Kuroda [500], molybdenum issorbed strongly on Amberlite CG 400 (Cl form) at pH 3 from seawater con-taining ascorbic acid, and is easily eluted with 6 M nitric acid. Molybdenum inthe effluent can be determined spectrophotometrically with potassium thio-cyanate and stannous chloride. The combined method allows selective andsensitive determination of traces of molybdenum in seawater. The precisionof the method is 2% at a molybdenum level of 10 µg/l. To evaluate the feasi-bility of this method, Kiriyama and Kuroda [500] spiked a known amount ofmolybdenum and analysed it by this procedure. The recoveries for 4 to 8 µgmolybdenum added to 500 or 1000 ml samples were between 90 and 100%.

Nutaksuka et al. [501] converted molybdenum to its molybdenum–phenyl-fluorone complex, then extracted the complex on a membrane filter prior tospectrophotometric determination on the membrane.

5.36 Molybdenum 187


A limited amount of work has been carried out on the determination ofmolybdenum in seawater by atomic absorption spectrometry and graphitefurnace atomic absorption spectrometry [137,502]. In a recommended proce-dure [503], a 50 ml sample of seawater at pH 2.5 is passed through a columnof 0.5 g p-aminobenzylcellulose, then the column is left in contact with 1 Mammonium carbonate for 3 h, after which three 5 ml fractions are collected.Finally, molybdenum is determined by atomic absorption at 313.2 nm usingthe hot graphite rod technique. At the 10 mg/l level, the standard deviation was0.13 µg.

Emerick [504] showed that sulfate interferes with the graphite furnaceatomic determination of molybdenum in aqueous solutions with concentra-tions of only 0.1% (w/v) sodium sulfate, causing complete elimination of themolybdenum absorbance peak in solutions free of other salts. Matrix modifi-cation with 0.5% (w/v) CaCl2·2H2O, in a volume equal to the sample, facilitatesthe determination of molybdenum in the presence of solutions containingas much as 0.4% (w/v) sodium sulfate. The need for matrix modification formolybdenum determination in natural waters appears to exist when sulfategreatly exceeds the equivalent calcium content. The routine use of a volume of0.5% CaCl2·2H2O equal to sample volume is recommended in the determina-tion of molybdenum in seawater.

Nakahara and Chakrabarti [137] showed that the seawater salt matrix can beremoved from the sample by selective volatilisation at 1700–1850 ◦C, but thepresence of sodium chloride, sodium sulfate, and potassium chloride causesa considerable decrease in molybdenum absorbance, and magnesium chlorideand calcium chloride cause a pronounced enhancement. The presence of mag-nesium chloride prevents the depressive effects. Samples of less than 50 µl canbe analysed directly without using a background corrector with a precisionof 10%.

These workers conclude that the selective volatilisation technique is highlysuitable for the determination of traces of molybdenum in synthetic (andmost probably real) seawater samples. It has the advantages of freedom fromcontamination and loss during sample preparation and is faster, and cheaper,than procedures using separations.

The sensitivity achieved should enable seawater samples to be analysed formolybdenum, because the concentration of molybdenum in seawater is usually2.1–18.8 µg/l. The selected temperature of 1700–1850 ◦C during the charringstage permits separation of the seawater matrix from the analyte prior toatomisation with the Perkin-Elmer Model 603 atomic absorption spectrometerequipped with a heated graphite atomiser (HGA-2100).

Kuroda et al. [505] determined traces of molybdenum in seawater by com-bined anion exchange and graphite-furnace atomic absorption spectrometry.


Trace amounts of molybdenum were concentrated from acidified seawateron a strongly basic anion exchange resin (Bio-Rad AG1 X-8 in the chlorideform) by treating the water with sodium azide. Molybdenum (VI) complexeswith azide were stripped from the resin by elution with ammonium chlo-ride/ammonium hydroxide solution (2 M/2 M). Relative standard deviationsof better than 8% at levels of 10 µg per litre were attained for seawater usinggraphite furnace atomic absorption spectrometry.


Specht and Beauchemin [506] have described an automated system to provideonline addition of isotopic spikes to seawater samples in the determination ofmolybdenum by inductively coupled plasma mass spectrometry.

5.36.4Electrochemical Methods

Hidalgo et al. [509] reported a method for the determination of molybde-num (VI) in natural waters based on differential pulse polarography. Thecatalytic wave caused by molybdenum (VI) in nitrate medium following pre-concentration by coflotation on ferric hydroxide was measured. For seawatersamples, hexadecyltrimethylammomum bromide with octadecylamine wasused as the surfactant. The method was applied to molybdenum in the range0.7–5.7 µg/l.

VandenBerg [510] carriedoutdirect determinationsofmolybdenumin sea-water by adsorption voltammetry. The method is based on complex formationof molybdenum (VI) with 8-hydroxyquinoline (oxine) on a hanging mercurydrop electrode. The reduction current of adsorbed complex ions was measuredby differential pulse adsorption voltammetry. The effects of variation of pHand oxine concentration and of the adsorption potential were examined. Themethod was accurate up to 300 nmol/l. The detection limit was 0.1 nmol/l.

Willie et al. [508] used linear sweep voltammetry for the determination ofmolybdenum. The molybdenum was adsorbed as the Eriochrome Blue BlackR complex on a static mercury drop electrode. The method was reported tohave a limit of detection of 0.50 µg/l and the results agreed well with certifiedvalues for two reference seawater samples.

Huaet al. [507]describedanautomatedmethod fordeterminationofmolyb-denum in seawater by means of constant-current reduction of the adsorbed 8-quinolinol complex in a computerised flow potentiometric stripping analyser.The complex was adsorbed onto a molybdenum film electrode at –0.2 V andstrippedat–0.42 V.Theauthors reportmeasuringmolybdenumat8.9±1.3 µg/lin reference seawater NASS-1, with a certified value of 11.5 ± 1.9 µg/l.

5.37 Neodymium 189

5.36.5X-ray Fluorescence Spectrometry

X-ray fluorescence was used for the determination of molybdenum in seawaterin a method described by Kimura et al. [511]. Molybdenum is coprecipitatedwith sodiumdiethyldithiocarbamate,which ismeasuredbyX-rayfluorescence.They report a detection limit of 0.3 µg/l and a relative standard deviationof 2.9%.

5.36.6Miscellaneous

Various other techniques have been used to determine molybdenum, includingadsorption voltammetry [510], electron-paramagnetic resonance spectrome-try [512], and neutron activation analysis [513, 514]. EPR spectrometry iscarried out on the isoamyl alcohol soluble Mo(SCN)5 complex and is capableof detecting 0.46 mg/l molybdenum in seawater. Neutron activation is carriedout on the β-naphthoin oxime [514] complex and the pyrrolidone dithiocar-bamate and diethyldithiocarbamate complex [513]. The neutron activationanalysis method [514] was capable of determining down to 0.32 µg/l of molyb-denum in seawater.

Monien et al. [515] have compared results obtained in the determinationof molybdenum in seawater by three methods based on inverse voltammetry,atomic absorption spectrometry, and X-ray fluorescence spectroscopy. Onlythe inverse voltammetric method can be applied without prior concentrationof molybdenum in the sample, and a sample volume of only 10 ml is adequate.Results of determinationsbyall threemethodsonwater samples fromtheBalticSea are reported, indicating their relative advantages with respect to reliability.

Shriadah et al. [516] determined molybdenum VI in seawater by densitom-etry after enrichment as the Tiron complex on a thin layer of anion exchangeresin. There were no interferences from trace elements or major constituentsof seawater, except for chromium and vanadium. These were reduced by theaddition of ascorbic acid. The concentration of dissolved molybdenum (VI) de-termined in Japanese seawater was 11.5 µg/l, with a relative standard deviationof 1.1%.

An adsorbing colloid formation method has been used to separate molyb-denum from seawater prior to its spectrophotometric determination by thethiocyanate procedure [517].

See also Sects. 5.74.5, 5.74.6, 5.74.8, 5.74.9, 5.74.12, 5.74.14, and 5.74.15.

5.37Neodymium

See Sect. 5.49.


5.38Neptunium

See Sect. 7.2.6.

5.39Nickel


The concentration of nickel in natural waters is so low that one or two en-richment steps are necessary before instrumental analysis. The most commonmethod is graphite furnace atomic absorption after preconcentration by sol-vent extraction [122] or coprecipitation [518]. Even though this technique hasbeen used successfully for the nickel analyses of seawater [519, 520] it is vul-nerable to contamination as a consequence of the several manipulation stepsand of the many reagents used during preconcentration.

Nickel has been determined spectrophotometrically in seawater in amountsdown to 0.5 µg/l as the dimethylglyoxime complex [521, 522]. In one proce-dure [521] dimethylglyoxime is added to a 750 ml sample and the pH adjustedto 9–10. The nickel complex is extracted into chloroform. After extraction into1 M hydrochloric acid, it is oxidised with aqueous bromine, adjusted to pH 10.4,and dimethylglyoxime reagent added. It is made up to 50 ml and the extinctionof the nickel complex measured at 442 nm. There is no serious interferencefrom iron, cobalt, copper, or zinc but manganese may cause low results.

In another procedure [522] the sample of seawater (0.5–3 litres) is filteredthrough a membrane-filter (pore size 0.7 µm) which is then wet-ashed. Thenickel is separated from the resulting solution by extraction as the dimethyl-glyoxime complex and is then determined by its catalysis of the reaction ofTiron and diphenylcarbazone with hydrogen peroxide, with spectrophoto-metric measurement at 413 nm. Cobalt is first separated as the 2-nitroso-1-naphthol complex, and is determined by its catalysis of the oxidation ofalizarin by hydrogen peroxide at pH 12.4. Sensitivities are 0.8 µg/l (nickel) and0.04 µg/l (cobalt).


Rampon and Cavelier [523] used atomic absorption spectrometry to deter-mine down to 0.5 µg/l nickel in seawater. Nickel is extracted into chloroformfrom seawater (500 ml) at pH 9–10, as its dimethylglyoxime complex. Severalextractions and a final washing of the aqueous phase with carbon tetrachloride

5.39 Nickel 191

are required for 100% recovery. The combined organic phases are evaporatedto dryness and the residue is dissolved in 5 ml of acid for atomic-absorptionanalysis.

Lee [524] described a method for the determination of nanogram or sub-nanogramamountsofnickel in seawater.Dissolvednickel is reducedby sodiumborohydride to its elemental form, which combines with carbon monoxideto form nickel carbonyl. The nickel carbonyl is stripped from solution bya helium–carbon monoxide mixed gas stream, collected in a liquid nitrogentrap, and atomised in a quartz tube burner of an atomic absorption spec-trophotometer. The sensitivity of the method is 0.05 ng of nickel. The precisionfor 3 ng nickel is about 4%. No interference by other elements is encounteredin this technique.

Between 0.3 and 0.6 µg/l nickel was found by this method, in a verticalprofile of water samples taken down to 1200 m in the Santa Catalina Basin.

Nishioka et al. [525] coprecipitated nickel from seawater with sodium di-ethyldithiocarbamate, filtered, and redissolved the precipitate with nitric acidfollowed by electrothermal atomic absorption spectrophotography determi-nation of the nickel. The detection limit was 0.5 µg/l and the relative standarddeviation was 13.2% at the 2 µg/l level.


Van den Berg and Nimmo [526] studied the complexation of nickel withdimethylglyoxime in seawater to determine nickel complexing capacities inseawater. Seawater samples were collected from the Menai Streets, LiverpoolBay, and the English Channel and used to test the speciation procedures. Thetheory for the determination of complexing capacities is presented. Seawatercontaining 0.01 M borate buffer and 0.0001 M dimethylglyoxime was pipet-ted into 10–15 separate Teflon voltammetric cells. Nickel was then added togive a concentration range between 1 and 20 nM. After equilibration, cathodicstripping voltammetry was used to determine the labile nickel concentra-tion by measuring the reduction current of nickel–dimethylglyoxime complexabsorbed on the hanging mercury drop electrode. Initial concentrations oftotal dissolved nickel were measured by cathodic stripping voltammetry with0.0001 M dimethylglyoxime after UV irradiation for 2 h. Values for nickel com-plexing capacities with dimethylglyoxime were determined for seawater ofseveral salinities by ligand competition with EDTA.

Donat and Bruland [217] determined low levels of nickel and cobalt inseawater by a voltammetric technique, and the nioxime complexes of the twoelements were concentrated on a hanging mercury drop electrode. The currentresulting from the reduction of Co (II) and Ni (II) was measured by differentialpulse cathodic stripping voltammetry. Detection limits are 6 pM (cobalt) and0.45 nM (nickel).


5.39.4Liquid Scintillation Counting

To determine 63Ni in seawater the nickel was adsorbed on to hydrous man-ganese dioxide and the precipitate dissolved in hydrochloric acid. The nickelwas then extracted with diethyldithiocarbamate in chloroform and determinedby liquid scintillation counting [527].

See also Sects. 5.74.4–5.74.6, and 5.74.8–5.74.17.

5.40Osmium

5.40.1Resonance Ionisation Mass Spectrometry

Koide et al. [528, 529] determined osmium in seawater by passing the waterdown an anion exchange resin column, followed by distillation of the osmiumtetroxide and detection by resonance ionization mass spectrometry.

5.41Palladium

Wang et al. [530] used a liquid membrane containing tri-N-octylamine toseparate palladium from seawater.

5.42Platinum


Platinum was determined in seawater by adsorptive cathodic stripping voltam-metry in a method described by Van den Berg and Jacinto [531]. The formazonecomplex is formed with formaldehyde, hydrazine, and sulfuric acid in the sea-water sample. The complex is adsorbed for 20 minutes at –0.925 V on thehanging mercury drop electrode. The detection limit is 0.04 pM platinum.

5.43Plutonium

See Sect. 7.2.1 and 7.2.7.

5.44 Polonium 193

5.44Polonium

See Sect. 7.1.2.

5.45Potassium

5.45.1Titration

Potentiometric titration has been applied to the determination of potassium inseawater [532–534]. Torbjoern and Jaguer [533–544] used a potassium selec-tive valinomycin electrode and a computerised semiautomatic titrator. Sam-ples were titrated with standard additions of aqueous potassium so that thepotassium to sodium ion ratio increased on addition of the titrant, and the con-tribution from sodium ions to the membrane potential could be neglected. Theinitial concentration of potassium ions was then derived by the extrapolationprocedure of Gran.

Marquis and Lebel [534] precipitated potassium from seawater or marinesediment pore water using sodium tetraphenylborate, after first removinghalogen ions with silver nitrate. Excess tetraphenylborate was then deter-mined by silver nitrate titration using a silver electrode for endpoint detection.The content of potassium in the sample was obtained from the difference be-tween the amount of tetraphenyl boron measured and the amount initiallyadded.

5.45.2Polarography

Polarography has also been applied to the determination of potassium in sea-water [535]. The sample (1 ml) is heated to 70 ◦C and treated with 0.1 M sodiumtetraphenylborate (1 ml). The precipitated potassium tetraphenylborate is fil-tered off, washed with 1% acetic acid, and dissolved in 5 ml acetone. Thissolution is treated with 3 ml 0.1 M thallium nitrate and 1.25 ml 2 M sodiumhydroxide, and the precipitate of thallium tetraphenylborate is filtered off. Thefiltrate is made up to 25 ml, and after de-aeration with nitrogen, unconsumedthallium is determined polarographically. There is no interference from 60 mgsodium, 0.2 mg calcium or magnesium, 20 µg barium, or 2.5 µg strontium.Standard eviations at concentrations of 375, 750, and 1125 µg potassium perml were 26.4, 26.9, and 30.5, respectively. Results agreed with those obtainedby flame photometry.



Ward [536] evaluated various types of potassium ion-selective electrodes forthe analysis of seawater. Three types of potassium ion-selective electrodeswere evaluated for their suitability for continuous monitoring and in situ mea-surement applications, in water of varying salinities and at temperatures of10 ◦C and 25 ◦C. The three types comprised a glass-membrane single elec-trode, a glass-membrane combination electrode, and a liquid-ion exchangeelectrode. Although all three electrode systems performed well in fresh water,the results obtained with the liquid-ion exchange electrode in seawater weresignificantly better than those with glass membranes. An accuracy of 5% couldbe achieved under certain conditions, but response times generally exceeded10 min, and glass-membrane electrodes were sensitive to external motion andflow variations.


5.46Praseodymium

See Sect. 5.49.2.

5.47Promethium

See Sect. 5.49.4.

5.48Radium

See Sect. 7.1.3.

5.49Rare Earths

These include the following 14 elements: cerium, praseodymium, neodymium,promethium, samarium, europium, gadolinium, terbium, dysprosium, holmi-um, erbium, thulium, ytterbium, and lutetium.

5.49.1Cerium

Shigematsu et al. [626] determined cerium fluorometrically at the 1 µg/l levelin seawater. Quadrivalent cerium is coprecipitated with ferric hydroxide and

5.49 Rare Earths 195

the precipitate is dissolved in hydrochloric acid; interfering ions are removedby extraction with isobutyl methyl ketone. The aqueous phase is evaporatedalmost to dryness with 70% perchloric acid, then diluted with water and passedthrough a column of bis-(2-ethylhexyl) phosphate on poly (vinyl chloride),from which CeIV is eluted with 0.3 M perchloric acid. The eluate is evaporated,then made 7 M in perchloric acid and treated with TiIII, and the resulting CeIII

is determined spectrofluorometrically at 350 nm (excitation at 255 nm).Cerium was included in a list of 14 elements determined by Lee et al. [627]

in seawater using neutron activation analysis. The metals were first precon-centraed on a mixture of Chelex 100 and glass powder. The elements weredesorbed from the column by 4 M nitric acid, and aqueous solution was irra-diated for 3 days and subjected to γ -ray spectrometry method with a Ge(Li)detector coupled to a 4000-channel analyser. Cerium was found to be present tothe extent of 16.7 µg/l in water taken from the Kwangyang Bay (South Korea).

For further details see Sect. 5.49.15.Lieser et al. [628] studied the application of neutron activation analysis to

the determination of trace elements in seawater. The rare earths included inthis study were cerium and europium. The element concerned were adsorbedonto charcoal. Between 75% and 100% of the elements were adsorbed onto thecharcoal which was then subjected to analysis by neutron activation analysis.Cerium (300 µg/l) and europium (0.00082 µg/l) were found in North Sea waterby this method.

For further discussion see Sect. 5.49.15.

5.49.2Praseodymium

See Sect. 5.49.15.

5.49.3Neodymium

See Sect. 5.49.15.

5.49.4Promethium

See Sect. 5.49.15.

5.49.5Samarium

See Sect. 5.49.15.


5.49.6Europium

See Sect. 5.49.15.

5.49.7Gadolinium

See Sect. 5.49.15.

5.49.8Terbium

See Sect. 5.49.15.

5.49.9Dysprosium

See Sect. 5.49.15.

5.49.10Holmium

See Sect. 5.49.15.

5.49.11Erbium

See Sect. 5.49.15.

5.49.12Thulium

See Sect. 5.49.15.

5.49.13Ytterbium

See Sect. 5.49.15.

5.49.14Lutetium

See Sect. 5.49.15.

5.49 Rare Earths 197

5.49.15Analysis of Rare Earth Mixtures

Elderfield and Greaves [629] have described a method for the mass spectromet-ric isotope dilution analysis of rare earth elements in seawater. In this method,the rare earth elements are concentrated from seawater by coprecipitation withferric hydroxide and separated from other elements and into groups for anal-ysis by anion exchange [630–635] using mixed solvents. Results for syntheticmixtures and standards show that the method is accurate and precise to ±1%;and blanks are low (e.g., 10–12 moles La and 10–14 moles Eu). The methodhas been applied to the determination of nine rare earth elements in a varietyof oceanographic samples. Results for North Atlantic Ocean water below themixed layer are (in 10–12 mol/kg) 13.0 La, 16.8 Ce, 12.8 Nd, 2.67 Sm, 0.644 Eu,3.41 Gd, 4.78 Dy, 407 Er, and 3.55 Yb, with enrichment of rare earth elements indeep ocean water by a factor of 2 for the light rare earth elements, and a factorof 1.3 for the heavy rare earth elements.

Elution volume calibrations were performed using radioactive tracers ofthe rare earth elements and 133Ba, with atomic-absorption or flame-emissionanalysis of iron, sodium, potassium, calcium, and magnesium. As shown inFig. 5.14, any barium added to the second columns is eluted at the start of the“light rare earth element fraction”. To ensure barium removal the sample canbe put through the first column again.

Ion exchange chromatography using Chelex 100 resin has been used forthe concentration of rare earth elements from large volumes of seawater, withrecoveries of 85–112% [636].

The chemistry of rare earth elements makes them particularly useful instudies of marine geochemistry [637]. But the determination of rare earthsin seawater at ultratrace levels has always been a difficult task. Of the variousmethodsapplied, instrumentalneutronactivationanalysis and isotopedilutionmass spectrometrywere themain techniquesused for thedeterminationof rareearths in seawater. However, sample preparation is tedious and large amountsof water are required in neutron activation analysis. In addition, the methodcan only offer relatively low sample throughputs and some rare earths cannotbedetermined.Themaindrawbacksof isotopicdilutionmass spectrometry arethat it is time-consuming and expensive, and monoisotopic elements cannotbe determined as well.

At present, inductively coupled plasma mass spectrometry providesa unique, powerful alternative for the determination of rare earths in nat-ural samples [638, 639]. Nevertheless, its application to the determinationof rare earths at ultratrace concentration level in seawater is limited, be-cause highly saline samples can cause both spectral interferences and ma-trix effects [640]. Therefore, a separation of the matrix components and pre-concentration of the analytes are prerequisites. To achieve this goal, manypreconcentration techniques have been used, including coprecipitation with


Figure 5.14. Elution curves of the rare earth elements in interfering elements for anion ex-change separation using CH3COOH/HNO3 (1st columns) and MeOH/HNO3 (2nd columns)mixtures. Source: [629]

iron hydroxide [641–643], ion exchange with Chelex 100 [636,644–646], treat-mentwith silica-immobilised8-hydroxyquinoline [647,648], solventextractionwith bis (2-ethylhexyl) hydrogen phosphate/2-ethylhexyl dihydrogen phos-phate [649,650], and sorption on activated carbon [651]. Coprecipitation withiron hydroxide requires an additional separation step, usually cation exchangechromatography, to remove magnesium, calcium, and iron before determina-tion. Chelation with Chelex 100 requires removal of calcium and magnesiumby careful washing with ammonium acetate or by cation and anion exchangechromatography prior to elution of rare earths. The solvent extraction tech-nique reported by Shabani et al. [649,650] involves many manipulations, suchas scrubbing, stripping from the organic phase, washing the aqueous solution,and evaporating the final solution. The concentration technique reported byEsser et al. [647] and Halicz et al. [648] can only concentrate rare earths atpH 8–9, and sorption on activated-carbon [651] is time-consuming.

Tian-Hong Zhang et al. [652] have reported a new ion exchange chelatingfibre with aminophosphonic and dithiocarbamate groups, based on polyacry-lonitrile for the preconcentration of rare earth elements in seawater prior totheir determination by inductively coupled plasma mass spectrometry. Rare

5.50 Rhenium 199

earths can be easily separated from the high-saline matrix using the fibre. Theoptimum experimental parameters, such as pH, flow rate, sample volume, andeffect of matrix ions for preconcentration of rare earths were investigated. Allrare earths in water can be quantitatively retained in the acidity range pH 3–6by the fibre and then eluted quantitatively with 0.01 mol/l ammonium citrate.The fibre has been applied to the concentration of rare earths in seawater. Therelative standard deviations for the determination of rare earths in seawaterat nanogram per liter levels were found to be less than 5%. Reasonably goodagreement is obtained with the data reported in the literature.

Wen et al. [950] used 8-hydroxyquinoline immobilised on a polyarylonitrilehollow fibre membrane to achieve a 300-fold concentration factor for rare earthelements in seawater.

See Sect. 5.74.16.

5.50Rhenium

Rhenium was one of the last stable elements to be discovered, one of the leastabundant metals in the earth’s crust, and one of the most important sentinels ofreducingaqueousenvironments through its abundance in sediments.Althoughits chemistry is fairly well understood, its marine chemistry is as yet poorlydeveloped. In addition, the understanding of rhenium’s marine chemistry willprovide an entry to the understanding of the marine chemistry of technetium,an element which is just above rhenium in group VIIA (group 7 in 1985notation) of the periodic table. Technetium has only unstable isotopes, whichoriginate primarily in nuclear weapon detonations and nuclear reactor wastes.These two elements have remarkably similar chemistries. Rhenium’s solutionchemistry primarily involves anionic species in the IV, V, and VIII oxidationstates. The oxo-anion perrhenate is especially stable.


Koide et al. [537] have described a graphite furnace atomic absorption methodfor the determination of rhenium at picomolar levels in seawater and parts-per-billion levels in marine sediments, based upon the isolation of heptavalentrhenium species upon anion exchange resins. All steps are followed with 186-rhenium as a yield tracer. A crucial part of the procedure is the separation ofrhenium from molybdenum, which significantly interferes with the graphitefurnace detection when the Mo:Re ratio is 2 or greater. The separation isaccomplished through an extraction of tetraphenylarsonium perrhenate intochloroform, in which the molybdenum remains in the aqueous phase.

It was observed by these workers that the rhenium signal was attenuatedby as little as 10 ng or less of molybdenum in the isolate. Thus, importance is


placed upon molybdenum decontamination steps. In seawaters as well as inmany marine sediments the Mo:Re ratio varies by about a factor of 1000. Inaddition, a clean separation of rhenium from other elements (the salt effect)is required. Otherwise, false peaks result upon atomisation, due to the highbackground generated by impurities.

The seawater concentration of rhenium is in the range from less than 3to 11 ng/l, compared to iridium, platinum, and gold, whose concentrationsusually do not exceed 0.3 ng/l.


MatthewsandRiley [538]havedescribed the followingprocedure fordetermin-ing down to 0.06 µg/l rhenium in seawater. From 6 to 8 µg/l rhenium was foundin Atlantic seawater. The rhenium in a 15 litre sample of seawater, acidified withhydrochloric acid, is concentrated by adsorption on a column of De-Acidite FFanion exchange resin (Cl– form), followed by elution with 4 M nitric acid andevaporation of the eluate. The residue (0.2 ml), together with standards andblanks, is irradiated in a thermal neutron flux of at least 3 ×1012 cm–2 s–1 for atleast 50 hours. After a decay period of 2 days, the sample solution and blank aretreated with potassium perrhenate as carrier and evaporated to dryness witha slight excess of sodium hydroxide. Each residue is dissolved in 5 M sodiumhydroxide. Hydroxylammonium chloride is added to reduce TcVII, which arisesfrom 99mTc from activation of molybdenum present in the samples, and theReVII is extracted selectively with ethyl methyl ketone. The extracts are evap-orated, the residue is dissolved in formic acid:hydrochloric acid (19:1), therhenium is adsorbed on a column of Dowex 1, and the column is washedwith the same acid mixture followed by water and 0.5 M hydrochloric acid; therhenium is eluted at 0 ◦C with acetone–hydrochloric acid (19:1), and is finallyisolated by precipitation as tetraphenylarsonium perrhenate. The precipitate isweighed to determine the chemical yield, and the 186Re activity is counted withan end-window Geiger–Müller tube. The irradiated standards are dissolved inwater together with potassium perrhenate. At a level of 0.057 µg/l rhenium, thecoefficient of variation was ±7%.

5.51Rubidium


Shen and Li [149] extracted rubidium (and caesium) from brine samples with4-tert-butyl-2-(α-methyl-benzyl) phenol prior to atomic absorption determi-nation of the metal.

5.52 Ruthenium 201

5.51.2Spectrometry

Schoenfeld and Held [539] used a spectrochemical method to determine ru-bidium in seawater. They determined concentrations of rubidium in the range0.008–0.04 µg/ml in the presence of varying proportions and concentrationsof other salts as internal standard. The coefficient of variation ranged from 7to 25% for simulated seawater standards.


Isotope dilution mass spectrometry has been used to determine traces ofrubidium in seawater [540].


Rubidiumhasbeendetermined inseawaterbyX-rayfluorescence spectrometrywith a detection limit of 2–4 µg of rubidium. The rubidium was coprecipitatedwith N3K2[Fe(CN6)]2 [541].

See also Sects. 5.74.4, 5.74.9 and 5.74.14.

5.52Ruthenium

See Sect. 7.2.8.

5.53Samarium

See Sect. 5.49.15.

5.54Scandium

See Sect. 5.74.15.

5.55Selenium

In recent years, the physiological role of selenium as a trace element has createdconsiderable speculation and some controversy. Selenium has been reported ashavingcarcinogenicaswell as toxicproperties; otherauthoritieshavepresented


evidence that selenium is highly beneficial as an essential nutrient [544,545]. Itssignificance and involvement in the marine biosphere is not known. A reviewof the marine literature indicates that selenium occurs in seawater as seleniteions (SeO2–

2 ) with a reported average of 0.2 µg/l [543].

5.55.1Spectrophotometry

Ferric hydroxide coprecipitation techniques are lengthy, two days being neededfor a complete precipitation. To speed up this analysis, Tzeng and Zeitlin [595]studied the applicability of an intrinsically rapid technique, namely adsorptioncolloid flotation. This separation procedure uses a surfactant–collector–inertgas system, in which a charged surface-inactive species is adsorbed on a hy-drophobic colloid collector of opposite charge. The colloid with the adsorbedspecies is floated to the surface with a suitable surfactant and inert gas, and thefoam layer is removed manually for analysis by a methylene blue spectrometricprocedure. The advantages of the method include a rapid separation, simpleequipment, and excellent recoveries. Tzeng and Zeitlin [595] used the floationunit that was devised by Kim and Zeitlin [517].

Other workers have described spectrophotometric methods [549–554].


Neve et al. [547] digested the sample with nitric acid. After digestion the sampleis reacted selectively with an aromatic o-diamine, and the reaction product isdetected by flameless atomic absorption spectrometry after the addition ofnickel (III) ions. The detection limit is 20 mg/l, and both selenium (IV) andtotal selenium can be determined. There was no significant interference ina saline environment with three times the salinity of seawater.


Cutter [548]has surveyed the applicationof this technique to thedeterminationof selenium in seawater.


Cathodic stripping voltammetry has been used to determine down to 20 ng/lselenium in seawater [542]. The selenium is extracted with 3,3′ diamino-benzidine and concentrated as HgSe on a mercury electrode at –0.45 V.

5.56 Silver 203

Certain trace substances such as selenium (IV) can be determined by differ-ential cathodic stripping voltammetry (DPCSV). For selenium a rather positivepreconcentration potential of –0.2 V is adjusted. Selenium (IV) is reduced toSe2–, and Hg from the electrode is oxidised to Hg2+ at this potential. It forms,with Se2– on the electrode, a layer of insoluble HgSe, and in this manner thepreconcentration is achieved. Subsequently the potential is altered in the ca-thodic direction in the differential pulse mode. The resulting mercury (II) peakproduced by the HgII reduction is proportional to the bulk concentration ofSeIV in the analyte.


Measures and Burton [556] used gas chromatography to determine seleniteand total selenium in seawater.

Shimoishi [555] determined selenium by gas chromatography with electroncapture detection. To 50–100 ml seawater was added 5 ml concentrated hy-drochloric acid and 2 ml 4-nitro-o-phenylenediamine (1%) and, after 2 hours,the product formed was extracted into 1 ml of toluene. The extract was washedwith 2 ml of 7.5 M hydrochloric acid, then a sample (5 µl) was injected intoa glass gas–liquid chromatography column (1 m × 4 mm) packed with 15%of SE-30 on Chromosorb W (60–80 mesh) and operated at 200 ◦C with nitro-gen (53 ml/min) as carrier gas. There is no interference from other substancespresent in seawater. The detection limit is 5 ng/l with 200 ml samples, and theprecision at a Se level of 0.025 µg/l is 6%.


This technique has been applied to the determination of selenium in seawa-ter [603, 604].


5.56Silver


Bermejo-Barrera et al. [557] have described an electrothermal atomic absorp-tion spectrometric method for the determination of silver at the ppb level inseawater.

Miller and Bruland [558] have described an equilibration/solvent extractionmethod based on competition for silver between sample ligands and added


diethyldithiocarbamate for thedeterminationofµg/l levels of silver in seawater.Detection was achieved by graphite furnace atomic absorption spectrometry.


Kawabuchi and Riley [559] used neutron activation analysis to determinesilver in seawater. Silver in a 4 l sample of seawater was concentrated by ionexchange on a column (6 cm× 0.8 cm) containing 2 g of DeAcidite FF-IP resin,previously treated with 50 ml of 0.1 M hydrochloric acid. The silver was elutedwith 20 ml of 0.4 M aqueous thiourea and the eluate was evaporated to dryness,transferred to a silica irradiation capsule, heated at 200 ◦C and ashed at 500 ◦C.After sealing, the capsule was irradiated for 24 hours in a thermal neutronflux of 3.5 ×1012 cm–2 s–1, and after a decay period of 2–3 days, the 110mAgarising from the reaction 199mAg (n, γ ) 110mAg was separated by a conventionalradiochemical procedure. The activity of the 110mAg was counted with an end-window Geiger–Müller tube, and the purity of the final precipitate was checkedwith a Ge(Li) detector coupled to a 4000-channel analyser. The method gavea coefficient of variation of ±10% at a level of 40 ng silver per litre.


5.57Sodium

5.57.1Amperometry

In the indirect amperometric method [560], saturated uranyl zinc acetatesolution is added to the sample containing 0.1–10 mg sodium. The solutionis heated for 30 minutes at 100 ◦C to complete precipitation. The solution isfiltered and the precipitate washed several times with 2 ml of the reagent andthen five times with 99% ethanol saturated with sodium uranyl zinc acetate.The precipitate is dissolved and diluted to a known volume. To an aliquotcontaining up to 1.7 mg zinc, 1 M tartaric acid (2–3 ml) and 3 M ammoniumacetate (8–10 ml) are added and the pH adjusted to 7.5–8.0 with 2 M aqueousammonia. The solution is diluted to 25 ml and an equal volume of ethanoladded. It is titrated amperometrically with 0.01 M K4Fe(CN)6 using a platinumelectrode. Uranium does not interfere with the determination of sodium.

5.57.2Polarimetry

In the indirect polarimetry method [561] sodium is precipitated as the zincuranyl acetate salt and the uranium present in the precipitate is determined

5.58 Strontium 205

polarimetrically after reaction with (+)-tartaric acid. The sample is diluted tocontain 0.1–1% (w/v) of sodium. A portion (1–2 ml) is treated with saturatedaqueous zinc uranyl acetate (10–20 ml) and the mixture evaporated to thevolume of reagent solution added. It is cooled then the precipitate is filteredoff and washed with reagent solution and with saturated ethanolic zinc uranylacetate. The precipitate is dissolved in water, 1 M tartaric acid (15 ml) is added,the pH adjusted to 5 with aqueous sodium hydroxide and diluted to 50 ml.The optical rotation is measured in a 20 cm tube and the sodium content ofthe sample determined by reference to a calibration graph, which is rectilinearover the range 1.62–16.2 mg sodium per 50 ml. The maximum error was 2.2%.

5.58Strontium


Carr [562] has studied the effects of salinity on the determination of strontiumin seawater by atomic absorption spectrometry using an air–acetylene flame.Using solutions containing 7.5 mg/l strontium and between 5 and 14% sodiumchloride, he demonstrated a decrease in absorption with increasing sodiumchloride concentration. To overcome this effect a standard additions procedureis recommended.


5.59Technetium

See Sect. 7.1.4.

5.60Tellurium


Petit [563] has described a method for the determination of tellurium in seawa-ter at picomolar concentrations. Tellurium (VI) was reduced to tellurium (IV)by boiling in 3 M hydrochloric acid. After preconcentration by coprecipitationwith magnesium hydroxide, tellurium was reduced to the hydride by sodiumborohydrate at 300 ◦C for 120 seconds, then 257 ◦C for 12 seconds. The hydridewas then measured by atomic absorption spectroscopy. Recovery was 90–95%and the detection limit was 0.5 pmol/l.


Andreae [564] coprecipitated tellurium (V) and tellurium (VI) from sea-water and other natural waters with magnesium hydroxide. After dissolutionof the precipitate with hydrochloric acid, the tellurium (IV) was reduced totellurium hydride in 3 M hydrochloric acid. The hydride was trapped inside thegraphite tube of a graphite furnace atomic absorption spectrometer, heatedto 300 ◦C, and tellurium (IV) determined. Tellurium (VI) was reduced to tel-lurium (IV) by boiling with hydrochloric acid and total tellurium determined.Tellurium (VI) was then calculated. The limit of detection was 0.5 pmol perlitre and precision 10–20%.


5.61Terbium

See Sect. 5.49.15.

5.62Thallium

See Sect. 5.74.5 and 5.74.8.

5.63Thorium

5.63.1Thermal Ion Mass Spectrometry

Moran et al. [565] and Guo et al. [566] have determined 230thorium and232thorium in seawater using, thermal ion mass spectrometry and secondaryion mass spectrometry, respectively.



Huh and Bacon [589] used neutron activation analysis to determine 232thoriumin seawater. Seawater samples were subjected to pre- and post-irradiation pro-cedures. Separation and purification of the isotopes, using ion exchange chro-matography and solvent extraction, were performed during pre-irradiation.After irradiation 233protactinium was extracted and counted. Yields were mon-itored with 230thorium and 231protactinium tracers. 232Thorium concentra-tions were 27 ×10–7 dpm/kg for deep water samples from below 400 m.

See Sect. 5.74.15.

5.64 Thulium 207

5.64Thulium

See Sect. 5.49.15.

5.65Tin


In an early method Kodama and Tsubota [567] determined tin in seawater byanion exchange chromatography and spectrophotometry with catechol violet.

After adjusting to 2 mol l–1 in hydrochloric acid, 500 ml of the sample is ad-sorbed on a column of Dowex 1-XS resin (Cl– form) and elution is then effectedwith 2 M nitric acid. The solution is evaporated to dryness after adding 1 M hy-drochloric acid, and the tin is again adsorbed on the same column. Tin is elutedwith 2 M nitric acid, and determined in the eluate by the spectrophotometriccatechol violet method. There is no interference from 0.1 mg of aluminium,manganese, nickel, copper, zinc, arsenic, cadmium, bismuth, or uranium; anytitanium, zirconium, or antimony are removed by ion exchange. Filtration ofthe sample through a Millipore filter does not affect the results, which are inagreement with those obtained by neutron activation analysis.


Dogan and Haerdi [568] and Bergerioux and Haerdi [569] determined totaltin in seawater by graphite furnace atomic absorption spectrometry. Theseworkers added 0.25 M 1,10-phenanthroline (0.1–1.0 ml) and 0.2 M tetraphenylboron(0.1–1.0 ml) (both reagentswere freshlyprepared) towater samples (50–1000 ml) which had been previously filtered through a 0.45 µm Millipore filter.The pH of this solution was adjusted to 5.0 before addition of coprecipitatingreagents. The precipitate thus obtained was either filtered or centrifuged anddissolved in a 1–5 ml portion of ammoniacal alcohol (methanol, ethanol oriso-propanol) solution of pH = 8–9 or in Lumatom®. For large volumes ofwater, the dissolution of the coprecipitate must be carried out with Lumatom®,since a precipitate is formed due to other ions present with ammoniacal alcoholsolution.


Brinckmann [570] used a gas chromographic method with or without hydridederivatisation for determining volatile organotin compounds (e.g., tetram-ethyltin) in seawater. For nonvolatile organotin compounds a direct liquid


chromatographic method was used. This system employs a Tenax GC poly-meric sorbent in an automatic purge and trap (P/T) sampler, coupled to a con-ventional glass column gas chromatograph equipped with a flame photomet-ric detector (FPD). Figure 5.15 is a schematic of the P/T-GC-FPD assemblywith typical operating conditions. Flame conditions in the FPD were tunedto permit maximum response to SnH emission in a H-rich plasma, as de-tected through narrow-bandpass interference filters (610 ± 5 nm) [571]. Twomodes of analysis were used: (1) volatile stannanes were trapped directly fromsparged 10–50 ml water samples with no pretreatment; and (2) volatilisedtin species were trapped from the same or replicate water samples follow-ing rapid injection of aqueous excess sodium borohydride solution directlyinto the P/T sparging vessel immediately prior to beginning the P/T cy-cle [572].

For either ion exchange resolution of aqueous cations, RnSn(4–n)+ aq [573]or their separation as ion pairs, [RnSn(4–n)+ X–4-n]0, on reverse bonded-phasecolumns [574] the method is restricted to “free” tin analytes. Unlike the vig-orous hydride derivatisation used in the gas chromatography–flame photo-metric detector method, common high-performance liquid chromatographysolvent combinations or their ionic addends will not usually provide sufficientcoordination strength to labilise organotin ions strongly bound to solids inenvironmental samples. Moreover, the high-performance liquid chromatogra-phy separations require that injected samples be free of particulates that mightclog the column or pumping system.

Brinckmann [570] generated calibration curves by the P/T gas chromatog-raphy–FPD method for borohydride reductions of SnIV, SnII, and Me2Sn2+

species to SnH4, SnH8, or SnMe2H2, respectively, in distilled water, 0.2 Msodium chloride, and bay water. All three analytes showed a substantial in-

Figure 5.15. The purge/trap GC-FPD system and operating conditions

5.65 Tin 209

crease in their calibration slopes in going from distilled water to 0.2 M sodiumchloride solution, the latter approximating the salinity and ionic strengthcommon to estuarine waters. Presumably these effects could arise from for-mation of chlorohydroxy tin species favouring more rapid hydridisation (see(5.1) and (5.2)) [575–577], as well as the more propitious partition coeffi-cients for dynamic gas stripping of the volatile tin hydrides from saline so-lutions [572]. In typical laboratory distilled water calibration solutions, only16% of SnII was recovered as SnH4, compared with SnIV, although this sen-sitivity ratio can probably be altered somewhat with pH changes [578, 579].However, in spiking anaerobic pre-purged Chesapeake Bay water with thesethree tin species, a striking reversal occurred in overall relative sensitivities,i.e., calibration slopes. Brinckmann [570] found that not only was Me2SnH2generation repressed by 50%, but very significantly, SnH4 formation fromSnIV was reduced by a factor of 15 as compared with the sodium chloridemedium.

RnSn4–n+(aq) + excess BH4– → RnSnH4–n (5.1)

4HSnO–2 + 3BH–

4 + 7H+ + H2O → 3H3BO3 + 4SnH4 (5.2)

The overall effect of estuarine water on the hydridisation process is thus oneof reducing yields of the three tin species tested. It is expected that not only thedissolved and particulate organics and chloride influence formation of Sn–Hbonds, but that other aquated metal ions play an important role, too. Severalworkers have reported that, for example, AsIII, AsV, CuII, CoII, NiII, HgII, PbII,and AgI interfere by unknown means at low concentrations [630, 633].

In summary, the hydride generation method cannot adequately differenti-ate between aquated SnIV and SnII, which may coexist in certain, especiallyanaerobic, environments found in marine waters. Inorganic tin, speciated as“tin (IV)”, should probably be regarded as “total reducible inorganic tin” untilmore discriminatory techniques become available [578, 580].


High-performance liquid chromatography, if coupled with a sensitive element-specificdetection systemsuchas atomicabsorption spectrometry, offers a valu-able tool for organotin speciation in complex fluids (especially for high organicloadings) not readily amenable to gas-phase derivatisation methods. In this ap-paratus the basic high-performance liquid chromatography setup described byBrinckmann [570] is coupled to graphite furnace atomic absorption spectrom-etry in a manner giving automatic periodic sampling (typically 45 s intervals)of the resolved eluents for tin-specific determination [574]. Injected sample


volumes may vary from 10 to 500 µl. Consequently, system sensitivity is broadand samples can be very representative.

Mixtures of R3S+n compounds (R = n-butyl, phenyl, cyclo-hexyl) were sep-

arated by ion exchange–high-performance liquid chromatography – graphitefurnace atomicabsorption spectrometry.The small spread incalibration slopessignifies similar efficiencies for their separation and column recovery, as well asgraphite furnace sensitivities. Considerably more sensitivity is possible withP/T-gas chromatography – flame photometric detector speciation of relatedorganotin species known to occur in environmental media [575, 578, 580].Much greater divergence in the P/T gas chromatography – flame photomet-ric detector system calibration slopes (ratios > 25) is obtained, probably asa result of different rates of hydride derivatisation during the fixed P/T purgetime (10 minutes), different partition coefficients affecting the rates at whichend species are sparged from the solution [572], or different retentivities onthe Tenax GC sorbent [577]. On the basis of the values obtained for the gaschromatograptic method with 10 ml sample volumes, nominal working rangesof 10–40 ng/l organotin are feasible.

Both systems are capable of at least a tenfold increase in sensitivity withonly minor changes in procedure and equipment. For high-performance liquidchromatography – graphite furnace atomic absorption spectrometry, this canbe achieved by both increasing injected sample size and opitimising flow rateswith a graphite furnace – atomic absorption spectrometry thermal programdesigned to give maximum atomisation efficiency for a specific organotinanalyte [575, 578]. For high-performance liquid chromatography – graphitefurnace atomic spectrometry, improvements are realised by adjusting purgeflow rate, and time while altering sodium borohydride additions to optimiseevolution of given organotin analyte [572]. Also, both increasing the samplevolumes [578, 580] and operating the Tenax GC trap at subambient tempera-tures [581–583] will yield lower working ranges.


A method described by Florence and Farrer [584] separated tin from its associ-ated lead by distillation from an aqueous sulfuric acid medium into which thevapour from boiling 50% hydrobromic acid is passed. The distillate providesan ideal supporting electrolyte for the determination of tin (II) (producedby reduction with hydrazinium hydroxide) by anodic stripping at a rotatingvitreous-carbon electrode in the presence of codeposited mercury [585, 586].The tin is deposited at –0.70 V versus the SCE for 5 minutes, and then strippedat –0.50 V during a sweep from –0.70 V to –0.45 V at 5 V per minute. Tin inseawater is coprecipitated on ferric hydroxide, and the precipitate is then dis-solved in the aqueous sulfuric acid, and subjected to the above procedure. Theaverage content for Pacific coastal waters was found to be 0.58 µg/l.

5.66 Titanium 211

5.65.6Miscellaneous

Gandon and Guegueniat [587] investigated the recovery of tin from seawaterusing manganese dioxide and using 125Sr as a tracer. Enrichment factors of 103

to 105 were found.Tao et al. [912] preconcentrated tin by derivatisation, followed by gas

chromatography – inductively coupled plasma mass spectrometry. Down to0.00001 ng/l tin could be determined in seawater.


5.66Titanium


Yang et al. [588] have described a spectrophotometric method for the determi-nation of dissolved titanium in seawater after preconcentration using sodiumdiethyldithiocarbamate. See also Sect. 5.74.14.

5.67Tungsten

Van den Sloot and Das [502] have described a method for the determinationof tungsten in seawater.

5.68Uranium


Agrawal et al. [590] determined down to 1 ppm of uranium in seawater byliquid extraction with N-phenyl-3-styrylacrylohydroxaminic acid followed byspectrophotometry.


Van den Berg and Huang [292] determined uranium (VI) in seawater by ca-thodic stripping voltammetry at pH 6.8 of uranium (VI)-catechol ions. A hang-ing mercury drop electrode was used. The detection limit was 0.3 nmol/l after


a collection period of 2.5 minutes. Interference by high concentrations ofiron (III) was overcome by selective adsorption of the uranium ions at a col-lection potential near the reduction potential of iron (III). Organic surfactantsreduced the peak heights for uranium by up to 75% at high concentrations.EDTA was used to eliminate competition by high concentrations of copper (II)for space on the surface of the drop.

Djogic et al. [591] determined down to 10 nM/l of uranium in seawaterusing square wave cathodic stripping voltammetry using a similar technique.Economou et al. [592] determined down to 0.1 µg/l of uranium (VI) in seawater.

5.68.3Polarography

Van den Berg and Nimmo [593] in their determination of uranium in seawateradded sample aliquots to the voltammetric cell of a polarograph together withbuffer comprising piperazine-N, N ′-bis (2-ethanesulfonic acid) monosodiumsalt and sodium hydroxide to give a final buffer concentration of 0.01 M.Oxine solution at 20 uM was also added. The effects of various ligands aschelating agents were investigated to determine the conditions allowing great-est sensitivity. Trans-1,2-diaminocyclohexane-N, N, N ′, N ′-tetra-acetic acid, 4–2 (2-pyridylazo) resorcinol, gallic acid, benzoin alpha-oxime, nitroso-R-salt,8-hydroxy-quinaldine, and dihydroxyanthraquinone did not give a peak foruranium in seawater at pH 6.9. 1,2-dihydroxybenzene-3,5-disulfonic acid, 3,2-dihydroxybenzoic acid, salicylaldoxime, 1-amino-2-naphthol-4-sulfonic acid,and cupferron did produce a peak for uranium. Best sensitivity for uraniumand lack of interferences occurred with 8-hydroxyquinoline. The procedurewas not possible for fresh waters because of the poor sensitivity of the com-parative method using catechol at low salinities.

5.68.4Miscellaneous

Spencer [594] has reviewed methods for the determination of uranium inseawater.

Hua et al. [595] have described an automated flow system for the constant-current reduction of uranium (VI) onto a mercury film-coated fibre electrode.Interference from iron (III) was eliminated by addition of sulfite. The resultsobtained for uranium (VI) in two reference seawater samples, NASS-1 andCASS-1, were 2.90 and 2.68 g/l, with standard deviations of 0.57 and 0.75 g/l,respectively.

See also Sects. 7.2.10, 5.74.9, 5.74.12, 5.74.14–5.74.16.

5.69 Vanadium 213

5.69Vanadium


Nishimura et al. [596] described a spectrophotometric method using 2-pyridylazoresorcinol for the determination of down to 0.025 µg/l vanadium in sea-water. The vanadium was determined as its complex with 4-(2-pyridylazo) re-sorcinol formed in the presence of 1,2-diaminocyclohexane-N, N, N ′, N ′-tetra-acetic acid. The complex was extracted into chloroform by coupling withzephiramine. Difficulties due to turbidity in the chloroform layer and incom-plete masking of some cations by 2-pyridylazoresorcinol were overcome byaddition of potassium cyanide and washing the chloroform layer with sodiumchloride solution. The extinction of the chloroform layer was measured at560 nm against water, as was that of a blank prepared with vanadium-freeartificial seawater. Sixteen foreign ions were investigated and no interferenceswere found at 5–100 times their usual concentration in seawater.

Kiriyama and Kuroda [597] combined ion exchange preconcentration withspectrophotometry using 2-pyridylazoresorcinol in the determination of vana-dium in seawater.

The sample (2 litres) diluted to 0.1 M in hydrochloric acid, filtered, and madeto 0.1 M in ammonium thiocyanate, is passed through a column of Dowex 1-X8resin (SCN form). The vanadium is retained and is eluted with concentratedhydrochloric acid. Thiocyanate in the eluate is decomposed by heating with ni-tric acid, and the solution is evaporated to fuming with sulfuric acid. A solutionof the residue is neutralised with aqueous ammonia and evaporated nearly todryness. The residue is treated with water and aqueous sodium hypobromite,and after 30 min with phenol, phosphate buffer solution of pH 6.5 and aqueous1,2-diaminocyclohexane-N, N, N ′, N ′-tetra-acetic acid, and the vanadium is de-termined spectrophotometrically at 545 nm with 4-(2-pyridylazo) resorcinol.Vanadium was determined in seawater at levels of 1.65 µg/l. After boiling suchsamples under reflux with potassium permanganate and sulfuric acid (to es-tablish the concentration of organically bound vanadium), values for vanadiumwere 30–60% higher than corresponding values obtained without oxidation.


Monien and Stangel [598] studied the performance of a number of alternativechelatingagents for vanadium, and their effect onvanadiumanalysis, by atomicabsorption spectrometry with volatilisation in a graphite furnace. Two promis-ing compounds were evaluated in detail, namely 4-(2-pyridylazo) resorcinol inconjunction with tetraphenylarsonium chloride and tetramethylenedithiocar-bamate. These substances, dissolved in chloroform, were used for extraction


of vanadium from seawater, and after concentrating the organic layer, 5 µl wasinjected into a pyrolytic graphite furnace coated with lanathanum carbide. Forboth reagents a linear concentration dependence was obtained between 0.5and 7 µg/l after extraction from a 100 ml sample.

Using the 2-pyridylazoresorcinol-tetraphenyl-arsonium chloride systema concentration of 1 µg/l could be determined with relative standard devia-tion of 7%.


Hastingsetal. [599]havedescribedamethodfor thedeterminationofpicogramquantities of vanadium in seawater by isotope dilution inductively coupledplasma mass spectrometry, with electrothermal vaporisation to introduce thesample into the plasma. A 50V isotope spike enriched to 44 atom% was equi-librated with samples, followed by chemical purification by cation exchangechromatography. Samples were introduced into the electrothermal vapourisa-tion unit with a palladium modifier and heated to 1000 ◦C. This quantitativelyeliminates the ClO+ isobaric interference with vanadium at m/z 51 for solutionsup to 0.5 M hydrochloric acid. The procedural blank was 0.27 pg of vanadium.Corrections for 50Ti and 50Cr, which interfere with the vanadium signal, weremade by measurement of 49Ti and 53Cr. These isobaric interferences and vari-able ArC levels were the limiting sources of error in the ID measurement,and diminished the detection limit to 6 pg of vanadium. The detection limitfor non-isotope dilution applications was 0.3 pg of vanadium in seawater. Ac-curacy was confirmed by determination of vanadium standards in calciumcarbonate, and by comparative measurement with ID thermal ionisation massspectrometry and graphite furnace atomic absorption spectroscopy.


Van den Berg and Huang [600] carried out direct electrochemical strippingof dissolved vanadium in seawater using cathodic stripping. Voltammetry wasperformed with a hanging mercury drop electrode. The detection limit was0.3 nmol/l after a collection period of 2 min.

Vega and Van den Berg [601] determined vanadium in seawater in amountsdown to 70 pM by absorptive stripping voltammetry.


Two methods for the determination of vanadium in seawater have been devel-oped which use neutron activation analysis and atomic absorption spectrome-

5.70 Ytterbium 215

try [602]. The solutions were left for 2–3 hours. These samples (1–3 litres) werepassed through a Dowex 1-X8 ion exchange column at a flow rate of 1.7 ml/min.The resin was then washed with 20 ml distilled water, and vanadium eluted with150 ml eluent solution.

The vanadium eluate was slowly evaporated under an infrared lamp, theresidue dissolved in 6 M hydrochloric acid (10 ml) containing 1 ml of the alu-minium chloride solution [603], and vanadium was determined by atomicabsorption spectrophotometry. For calibration, suitable standard solutionswere aspirated before and after each batch of samples.

The average concentration and standard deviation of the Pacific Ocean wa-ters (µg/l) were 2.00 ± 0.09 by neutron activation analysis, and 1.86 ± 0.12 byatomic absorption spectrometry. For the Adriatic water the corresponding val-ueswere about 1.7 µg/l. Thedifferencebetween thevalues for the sameseawateris within the range to be expected from the standard deviations observed.

Although the neutron activation analysis is inherently more sensitive thanthe atomic absorption spectrometry, both procedures yield a reliable measure-ment of vanadium in seawater at the natural levels of concentration.

See also Sects. 5.74.5, 5.74.8, 5.74.9, 5.74.12, and 5.74.14–5.74.17.

5.70Ytterbium

See Sect. 5.49.15.

5.71Yttrium

See Sect. 5.74.14.

5.72Zinc

Zinc has only been measured accurately in open ocean by a few investiga-tors [239, 604–607]. Few data are available because of very low zinc concen-trations in seawater and the ubiquitous sources of zinc contamination. Theuncertainty of all zinc measurements prior to these investigations, and thepaucity of reliable data since, have left little information for the environmentalchemist to unravel the biogeochemical behaviour of zinc or to detect watersperturbed by anthropogenic inputs.

Interest in zinc concentrations in the ocean stems from its dual role asa required nanonutrient and as a potential toxicant due to its widespreadindustrial and marine usage [519, 605]. The major inputs of zinc to surfaceseawater include atmospheric deposition (both natural and anthropogenic in


origin), fluvial run-off, and upwelled waters. Zinc exists at natural levels inNorth Pacific surface water at a total concentration of approximately 0.1 nM,increasing to 3 nM at 500 m, and reaching a maximum of ∼9 nM at depthsgreater than 2000 m [664]. Our present understanding of the behaviour of zincin themarineenvironment isbasedononlya fewverticalprofiles.Theseprofilesindicate that zinc is actively incorporated into phytoplankton in surface watersand transported to depth in association with particulate organic matter orpassively adsorbed onto particles. There is a high correlation between zinc anddissolved orthosilicic acid which indicates that zinc, like silicate, is regenerateddeep in the water column and has a long deep water residence time on the orderof 10 000 years [508,587]. Recent studies indicate that the majority of dissolvedzinc in seawater is organically complexed, but the origin and behaviour of thezinc-binding ligands have not been characterized [185, 608].


Nowicki et al. [609] have described a sensitive technique for the shipboard de-termination of zinc in seawater. The technique couples flow injection analysiswith fluorometric detection. A cation exchange column was used to separatezinc from interfering alkali and alkaline earth ions and to concentrate zincfrom seawater. The organic indicator ligand, ρ-tosyl-8-aminoquinoline, wasused to form a complex with zinc, the fluorescence of which was determinedwith a flow-through fluorometer. The detection limit (defined as three timesthe standard deviation of the blank, n = 4) was 0.1 nM for a 4.4 ml sample.The precision based on the replicate analysis of samples containing 4.3 nM Znwas ±6% (n = 5). A single sample can be analysed in 6 min. The techniquewas determined to be accurate on the basis of analysis of the standard seawa-ter solutions CASS-2 and NASS-2, and by comparison with previous reliableinvestigations. A typical profile of 12 samples along with standards and blankscan be completed in triplicate in 5.5 hours.


Graphite furnace atomic absorption spectrometry has also been used to deter-mine zinc [610, 611] in seawater with a detection limit of 2 µg [611]. Guevre-mont [610] has discussed the use of organic matrix modifiers for the directdetermination of zinc.

Dissolved zinc concentrations in seawater have been determined by precon-centration using organic extraction (using APDC/DDDC) or chelating resins(using Chelex 100), followed by graphite furnace atomic absorption spectrom-etry [122, 612–615] or isotope dilution mass spectrometry [612]. These pro-

5.72 Zinc 217

cedures must be performed in shore-based ultra-clean laboratories by highlytrained personnel.

Huang and Shih [616] used a graphite furnace atomic absorption spectrom-eter with a stabilised platform furnace involving atomisation from a graphitesurface pretreated with vanadium to determine down to 24 ppt of zinc in sea-water.

Akatsuka et al. [617] determined down to 2.4 ng/dm3 of zinc in seawater(500 ml sample) by preconcentration on a column of C18 resin coated withmethyltricapryl ammonium chloride, followed by graphite furnace atomicabsorption spectrometry.


A method described by Hirata and Honda [618] uses a flow injection analysismanifold for pH adjustment of a seawater sample, followed by concentrationof zinc on a column packed with Chelex 100 resin. The zinc was eluted withnitric acid and determined by atomic absorption spectrometry. The detectionlimit is 0.5 µg/l and the relative standard deviation is 2.7% at the 10 µg/l level.

5.72.4Stripping Voltammetry

Van den Berg [619] determined zinc complexing capacity in seawater by ca-thodic stripping voltammetry of zinc–ammonium pyrrolidine dithiocarba-mate complex ions. The successful application of cathodic stripping voltam-metry, preceded by adsorptive collection of complexes with ammonium pyrro-lidine dithiocarbamate for the determination of zinc complexing capability inseawater isdescribed.Thereductionpeakofzincwasdepressedasaresultof lig-and competition by natural organic material in the sample. Sufficient time wasallowed for equilibrium to occur between the natural organic matter and addedammonium pyrrolidine dithiocarbamate. Investigations of electrochemicallyreversible and irreversible complexes in seawater of several salinities are de-tailed, together with experimental measurements of ligand concentrations andconditional stability constants for complexing ligands. Results were compara-ble with those obtained by other equilibrium techniques, but the above methodhad a greater sensitivity.

Van den Berg [620] also reported a direct determination of sub-nanomolarlevels of zinc in seawater by cathodic stripping voltammetry. The ability of am-monium pyrrolidine dithiocarbamate to produce a significant reduction peakin the presence of low concentrations of zinc was used to develop a methodcapable of achieving levels two orders of magnitude below those achieved withanodic stripping voltammetry. Interference from nickel and cobalt ions couldbe overcome by using a collection potential of 1.3 V, and interference from


complexing material by ultraviolet irradiation. Zinc could be determined inseawater and fresh water. Zinc and nickel could be determined simultane-ously by using dimethylgloxime at a collection potential of –0.7 V, followedby ammonium pyrrolidine dithiocarbamate at –1.3 V. The sensitivity for thisdetermination was 3 pmol/l.

Zima and Van den Berg [621] determined zinc in seawater in amounts downto 3 nmol by cathodic stripping voltammetry.

Muzzarelli and Sipos [622] showed that a column of chitosan (15 × 10 mm)can be used to concentrate zinc from 3 litres of seawater before determinationby anodic-stripping voltammetry with a composite mercury–graphite elec-trode. Zinc (and lead) are eluted from the column by 2 M ammonium acetate(50 ml), copper by 0.01 M EDTA (10 ml), and cadmium by 0.1 M potassiumcyanide (3 ml).

Anodic stripping voltammetry using a tubular mercury-graphite electrode[623] has been employed to determine zinc in seawater. Zinc concentrations of1 ×109 M can be detected within 5 min using this system.

Analysis of total zinc by anodic stripping voltammetry is problematic be-cause of interference by the hydrogen wave in acidified samples, and due to theinability to detect organically complexed zinc at natural pH values near 8 [185].An improved understanding of zinc in marine systems now requires rapid, sen-sitive analytical methods that are less prone to contamination, and that can beperformed at sea [624].

5.72.5Miscellaneous

Adsorption colloid floation using dodecylamine as surfactant has been usedto separate zinc with 95% efficiency from seawater [625].

See also Sects. 5.74.4–5.74.6, 5.74.8–5.74.11, and 5.74.14–5.74.16.

5.73Zirconium

See Sects. 5.74.14 and 5.74.15.

5.74Multication Analysis

5.74.1Titration Procedures

Calcium and Magnesium

Mascini [653] described a potentiometric titration procedure using an OrionCu2+ state electrode for the determination of calcium and magnesium in sea-water.

5.74 Multication Analysis 219

The sample was mixed with an equal volume of borate buffer (pH 9.2). Titra-tion with 0.01 M EDTA gave two breaks corresponding to the concentration ofeach cation.

Jagner and Kerstein [654, 655] used computer-controlled high-precisioncomplexiometric titration for the determination of the total alkaline earthmetal concentration in seawater. Total alkaline earths were determined byphotometric titration using EDTA with eriochrome Black as indicator. Themethod yielded 63.32 µmole kg–1 for the total alkaline earth concentration instandard seawater of 3.5% salinity. The precision was about 0.01%.

5.74.2Spectrophotometric Procedure

Calcium, Magnesium, and Strontium

Pausch and Margerum [656] have described a differential kinetic method usingstopped-flow spectrophotometry.

Atienza et al. [657] reviewed the applications of flow injection analysiscoupled to spectrophotometry in the analysis of seawater. The method is basedon the differing reaction rates of the metal complexes with 1,2-diaminocycl-ohexane-N, N, N ′, N ′-tetra-acetate at 25 ◦C. A slight excess of EDTA is added tothe sample solution, the pH is adjusted to ensure complete formation of thecomplexes, and a large excess of 0.3 mM to 6 mM-Pb2+ in 0.5 M sodium acetateis then added. The rate of appearance of the PbII-EDTA complex is followedspectrophotometrically, 3 to 6 stopped-flow reactions being run in succession.Because each of the alkaline-earth–metal complexes reacts at a different rate,variations of the time-scan indicates which ions are present.

5.74.3Molecular Photoluminescence Spectrometry

Antimony and Arsenic

Tao et al. [658] have described a procedure in which antimony and arsenicwere generated as hydrides and irradiated with ultraviolet light. The broadcontinuous emission bands were observed in the ranges about 240–750 nmand 220–720 nm, and the detection limits were 0.6 ng and 9.0 ng for antimonyand arsenic, respectively. Some characteristics of the photoluminescence phe-nomenon were made clear from spectroscopic observations. The method wassuccessfully applied to the determination of antimony in river water and sea-water. The apparatus used in this technique is illustrated in Fig. 5.16.

Negative interferences by transition metal cations such as nickel and cop-per and nitrite were observed. However, these interferences have also beenreported for the hydride generation atomic absorption method, and are due to


Figure 5.16. Hydride generation system for photoluminescence. Source: [658]

the inhibition of hydride generation. There was no interference from volatileorganic compounds. Nitrate enhanced the luminescence signal. Antimony wasfound to occur in seawater at a level of 0.35 ng/ml.

5.74.4Flame Atomic Absorption Spectrometry

In general, this technique does not have adequate sensitivity for the deter-mination of the low levels of cations likely to occur in seawater. Couplingthe technique with a preconcentration procedure can, however, enable someanalyses to be carried out at the µg/l level.

Heavy Metals(Copper, Zinc, Lead, Cadmium, Iron, Magnesium, Nickel, Cobalt, and Silver)

Armannsson [659] has described a procedure involving dithizone extractionand flame atomic absorption spectrometry for the determination of cadmium,zinc, lead, copper, nickel, cobalt, and silver in seawater. In this procedure500 ml of seawater taken in a plastic container is exposed to a 1000 W mercuryarc lamp for 5–15 h to break down metal organic complexes. The solution isadjusted to pH 8, and 10 ml of 0.2% dithizone in chloroform added. The 10 mlof chloroform is run off and after adjustment to pH 9.5 the aqueous phase isextracted with a further 10 ml of dithizone. The combined extracts are washedwith 50 ml of dilute ammonia. To the organic phases is added 50 ml of 0.2 M-hydrochloric acid. The phases are separated and the aqueous portion washedwith 5 ml of chloroform. The aqueous portion is evaporated to dryness and theresidue dissolved in 5 ml of 2 M hydrochloric acid (solution A). Perchloric acid(3 ml) is added to the organic portion, evaporated to dryness, and a further2 ml of 60% perchloric acid added to ensure that all organic matter has been


oxidised. After evaporation, the sides of the beaker are washed down withaproximately 10 ml of distilled water, evaporated to dryness, and then theresidue is taken up in 5 ml of 2 M hydrochloric acid (solution B).

Cadmium is determined at 228.8 nm, zinc at 213.8 nm, and lead at 217.0 nmin solution A. In solution B, copper is determined at 324.7 nm, nickel at232.0 nm, cobalt at 240.7 nm, and silver at 328.1 nm, all using suitable scaleexpansion.

Detection limits achieved for the seven metals are between 0.04 µg/l (cobaltand lead) and 0.6 µg/l (zinc).

Olsen et al. [660] used a simple flow injection system, the FIAstar unit, toinject samples of seawater into a flame atomic absorption instrument, allowingthe determination of cadmium, lead, copper, and zinc at the parts per millionlevel at a rate of 180–250 samples per hour. Further, online flow injectionanalysis preconcentration methods were developed using a microcolumn ofChelex 100 resin, allowing the determination of lead at concentrations aslow as 10 µg/l, and of cadmium and zinc at 1 µg/l. The sampling rate wasbetween 30 and 60 samples per hour, and the readout was available within60–100 seconds after sample injection. The sampling frequency depended onthe preconcentration required.

Fang et al. [661] have described a flow injection system with online ionexchange preconcentration on dual columns for the determination of traceamounts of heavy metal at µg/l and sub-µg/l levels by flame atomic absorptionspectrometry (Fig. 5.17). The degree of preconcentration ranges from a factorof 50 to 105 for different elements, at a sampling frequency of 60 samples perhour. The detection limits for copper, zinc, lead, and cadmium are 0.07, 0.03,0.5, and 0.05 µg/l, respectively. Relative standard deviations are 1.2–3.2% atµg/l levels. The behaviour of the various chelating exchangers used was studiedwith respect to their preconcentration characteristics, with special emphasison interferences encountered in the analysis of seawater.

The flow injection AAS system with online preconcentration will challengethe position of the graphite furnace technique, because it yields comparablesensitivity at much lower cost by using simpler apparatus and separationmode.The method offers unusual advantages when matrices with high salt content(e.g., seawater) are analysed, because the matrix components do not reach thenebuliser.

Cabezon et al. [662] simultaneously separated copper, cadmium, and cobaltfrom seawater by coflotation with octadecylamine and ferric hydroxide ascollectors prior to analysis of these elements by flame atomic absorption spec-trometry. The substrates were dissolved in an acidified mixture of ethanol,water, and methyl isobutyl ketone to increase the sensitivity of the deter-mination of these elements by flame atomic absorption spectrophotometry.The results were compared with those of the usual ammonium pyrrolidinedithiocarbamate/methyl isobutyl ketone extraction method. While the meanrecoveries were lower, they were nevertheless considered satisfactory.


Figure5.17. Dual-column ion exchange preconcentration valve. SA, SB samples A and B; CA,CB ion exchange columns A and B; EA, EB eluant (2 M nitric acid) for columns A and B;WA, WB waste lines for samples and eluants A and B; W waste lines; AAS atomic absorptionspectrometer. The dimensions of the base plate of the valve are 70 × 45 × 10 mm. See textfor details of operation. Source: [661]

Zhuang et al. [664] used palladium salts as a coprecipitation carrier for theconcentration of cadmium, cobalt, and lead in seawater prior to analysis byatomic absorption spectrometry.

Jin [666] used ammonium pyrrolidine dithiocarbamate and electrothermalatomic absorption spectrometry to determine lead, cadmium, copper, cobalt,tin, and molybdenum in seawater.

Rodionova and Ivanov [667] used chelate extraction in the determinationof copper, bismuth, lead, cadmium, and zinc in seawater. The metal complexesof diethyl and dithiophosphates are extracted in carbon tetrachloride prior todetermination by atomic absorption spectrometry.

Chakraborti et al. [665] determined cadmium, cobalt, copper, iron, nickel,and lead in seawater by chelation with diethyldithiocarbamate from a 500 mlsample, extraction into carbon tetrachloride, evaporation to dryness, and re-dissolution in nitric acid prior to determination by electrothermal atomicabsorption spectrometry in amounts ranging from 10 pg (cadmium) to 250 pg(nickel).

Tony et al. [951] have discussed an online preconcentration flame atomicabsorption spectrometry method for determining iron, cobalt, nickel, magne-sium, and zinc in seawater. A sampling rate of 30 samples per hour was achievedand detection limits were 4.0, 1.0, 1.0, 0.5, and 0.5 µg/l, for iron, cobalt, nickel,magnesium, and zinc, respectively.


Silver, Cadmium, Lead, Cobalt, and Nickel

Cimadevilla et al. [691] compared wall, platform, and graphite furnace probeatomisation techniques in electrothermal atomic absorption spectrometry forthe determination of µg/l levels of silver, cadmium, and lead in seawater.

Chang et al. [952] used a miniature column packed with a chelating resinand an automatic online preconcentration system for electrothermal atomicabsorptionspectrometry todeterminecadmium,cobalt, andnickel in seawater.Detection limits of 0.12, 7 and 35 ng/l were achieved for cadmium, cobalt, andnickel, respectively, with very small sample volume required (400–1800 µl).

Potassium, Lithium, and Rubidium

Orren [663] used atomic absorption spectrometry to determine these ele-ments in seawater in both their soluble and insoluble forms. The alkali metalsare determined directly, but the other elements are first concentrated by sol-vent extraction. The particulate matter content is derived by dissolving themembranes used to filter the sample and determine the metals in the resultingsolution. For organic standards of known metal content, the efficiency of thetechnique was almost 100%.


Heavy Metals

The preponderance of work on multielement analysis in seawaters has beencarried out using the graphite furnace technique, as this has the additionalsensitivity over the direct technique that is required in seawater analysis.

Theoretical treatments of graphite furnace atomic absorption spectrome-try include a study of background signals due to sea salts [668], pyrometricmeasurement of furnace temperature [669], and methods of introducing thesample into the furnace [670].

Bengtsson et al. [671] found that the high background adsorption of so-lutions of trace metals containing up to 400 mg/l can be easily minimised byaddition of 2% v/v nitric acid. Of the several agents added in an attempt toeliminate the decrease in sensitivity caused by the salt and the variability insensitivity between graphite tubes, only lanthanum added at 1 g/l was effectivefor both lead and cadmium.

Campbell and Ottaway [672] have described a simple and rapid method forthe determination of cadmium and zinc in seawater, using atomic absorptionspectrometry with carbon furnace atomisation. Samples, diluted 1 + 1 withdeionised water, are injected into the carbon furnace and atomised in an HGA-72 furnace atomiser under gas-stop conditions. A low atomisation temperature


of 1492 ◦C is used to separate the atomic absorption signals from backgroundabsorption. Detection limits (2 SD) of 0.04 µg/l for cadmium and 1.7 µg/l forzinc are reported. These limits appear to be adequate for all but the cleanestseawater samples. The use of standard additions is essential because of theinterference from magnesium chloride, and also when samples of varyingsalinity have to be analysed.

Lead, cobalt, and nickel have been determined in seawater by atomic ab-sorption spectrometry after electrodeposition on pyrolytic graphite-coatedtubes [390]. The tubular, pyrolytic graphite-coated furnace has been incorpo-rated in a flow-through cell for the electrodeposition with mercury of heavymetals from seawater. After plating, the furnace is transferred to an atomicabsorption spectrometer for atomisation of the deposited metals. The flowassembly was tested for the analysis of lead in seawater, comparing results withthose obtained by anodic stripping voltammetry. The technique is appliedto the determination in seawater of both labile and total cobalt and nickel.These metals are irreversibly deposited on graphite and have poor sensitiv-ity using anodic stripping voltammetry, but are readily measured by atomicabsorption spectrometry. Measurements are reproducible with a relative stan-dard deviation of 15%. For 15- and 10-minute depositions, copper and nickelcharacteristic concentrations are 0.02 µg/l.

Stein et al. [673] have described a simplified, sensitive, and rapid method fordetermining low concentrations of cadmium, lead, and chromium in estuarinewaters. To minimise matrix interferences, nitric acid and ammonium nitrateare added for cadmium and lead; only nitric acid is added for chromium.Then 10, 20, or 50 µl of the sample or standard (the amount depending on thesensitivity required) is injected into a heated graphite atomiser, and specificatomic absorbance is measured. Analyte concentrations are calculated fromcalibration curves for standard solutions in demineralised water for chromium,or an artificial seawater medium for lead and cadmium.

Detection limits (µg/l) were 0.1 for cadmium, 4 for lead, and 0.2 for chromi-um. The relative standard deviations (n = 10) were 20, 9.5, and 18, respectively.

A graphite furnace procedure has been described [674] for the direct deter-mination of iron, chromium, and manganese in seawater and estuarine watersin which the interference normally associated with the presence of sodiumchloride is eliminated. The technique requires only very small sample volumes(10–20 µl) for the atomisation stage. The reproducility of the method was verygood. Sensitivities of 0.4, 0.2, and 0.07 µg/l and precisions of determination of4.5, 3, and 11% (at 2 µg/l level) were obtained for iron, manganese, and zinc.

Montgomery and Peterson [675] showed that ammonium nitrate used asa matrix modifier in seawater analysis to eliminate the interference of sodiumchloride degrades the pyrolytic coating on graphite-furnace tubes. The initiallyenhanced sensitivities for copper, manganese, and iron are maintained for upto 15 atomisations. There is then a rapid decline to a constant lower sensitivity.The characteristics depend strongly on the particular lot of furnace tubes. To


decrease the sodium chloride interference without using a matrix modifier, es-tuarine samples must be diluted (1 + 1) with pure water. Blanks and standardsare prepared and diluted with sample water containing low amounts of tracemetals to match the sample matrix.

Iron and manganese have been determined in saline pore water [676] by thefollowing technique.

Pre-acidifiedporewater (100 µl, dilutedwithMilliporeQ-water if necessary)was transferred, using an Eppendorf pipette, into a 10 ml volumetric Pyrexflask. To this flask nitric acid (50 µl) was added, and the solution was thenbrought to volume with Millipore Q-water. Standards were made up by addingvarious amounts to stock metal solutions (1 mg/l), nitric acid (50 µl), anda seawater solution (100 µl) of approximately the same salinity as the samples tobe analysed. This final addition ensures that the standards are of approximatelythe same ionic strength and contain the same salts as the samples.

The samples were analysed by injecting 25 µl aliquots into an HGA 2000Perkin-Elmer graphite furnace attached to a Jarrell–Ash 82–800 double beamatomic absorption spectrophotometer. Graphite tubes in the furnace were re-placed after 75–100 analyses. Metal concentrations were determined by com-paring the peak heights of the samples to the standard curve established bythe determination of at least five known standards. The detection limits ofthis technique for 1% absorption were 0.9 µmol/l (Fe), and 0.2 µmol/l (Mn).The coefficient of variation was: ±11% at 6.5 µmol/l for iron and +12% at11.8 µmol/l for manganese.

Cadmium, Copper, and Silver

Cadmium, copper, and silver have been determined by an ammonium pyrro-lidine dithiocarbamate chelation, followed by a methyl isobutyl ketone extrac-tion of the metal chelate from the aqueous phase [677], and finally followed bygraphite furnace atomic absorption spectrometry. The detection limits of thistechnique for 1% absorption were 0.03 µmol/l (copper), 2 nmol/l (cadmium),and 2 nmol/l (silver).

Yates [678] has discussed the application of graphite furnace atomic ab-sorption spectrometry to the determination of cadmium, copper, lead, nickel,and zinc in filtered saline water samples. He concludes that the determi-nation of these elements is possible with good precision and accuracy byflame or graphite furnace methods after ozonisation and matrix separationon Chelex 100 chelating resin. The limits of detection range between 0.01 µg/l(cadmium) and 50 µg/l (nickel). While application of direct graphite furnaceatomic absorption spectrometry (i.e., without Chelex 100 preconcentration)appears to be feasible for the determination of copper and manganese, it doesnot appear to be so for cadmium and lead, due to a large and variable suppres-sive interference, and it is concluded that considerable effort would be neededto develop a rapid procedure suitable for routine analysis.


Boyle and Edmond [679] determined copper, nickel, and cadmium in100 ml of seawater by coprecipitation with cobalt pyrrolidine dithiocarba-mate and graphite atomiser atomic absorption spectrometry. Concentrationranges likely to be encountered and estimated analytical precisions (1σ) are1–6 nmol/kg (±0.1) for copper, 3–12 nmol/kg (±0.3) for nickel, and 0.0–1.1 nmol/kg (±0.1) for cadmium.

Levels of copper, nickel, and cadmium found in the Sargosso Sea were 1.3–4,4.5–8.2, and 0.31–1.29 nmol/kg, respectively.

Ammonium pyrrolidine dithiocarbamate (APDC) chelate coprecipitationcoupled with flameless atomic absorption provides a simple and precise meth-od for thedeterminationofnanomolkg–1 levels of copper, nickel, andcadmiumin seawater. With practice, the method is not overly time-consuming. It is rea-sonable to expect to complete sample concentration in less than 20 min, diges-tion in about 4 h, and sample preparation in another hour. Atomic absorptiontime should average about 5 min per element. Excellent results have been ob-tainedon thedistributionofnickel andcadmiumin theoceanby this technique.

Brugmann et al. [680] compared three methods for the determination ofcopper, cadmium, lead, nickel, and zinc in North Sea and northeast Atlanticwaters. Two methods consisted of atomic absorption spectroscopy but withpreconcentration using either freon or methyl isobutyl ketone, and anodicstripping voltammetry was used for cadmium, copper, and lead only. Inexpli-cable discrepancies were found in almost all cases. The exceptions were thecadmium results by the two atomic absorption spectrometric methods, and thelead results from the freon with atomic absorption spectrometry and anodicscanning voltammetric methods.

The precision of the determinations is generally best using the freon ex-traction. This is probably because these extractions and determinations wereperformed under full clean-room conditions. The drawback of acid leachingfrom containers during long storage is small compared with the advantagesgained from working under clean-room conditions.

Bruland et al. [122] have shown that seawater samples collected by a varietyof clean sampling techniques yielded consistent results for copper, cadmium,zinc, and nickel, which implies that representative uncontaminated sampleswere obtained. A dithiocarbamate extraction method coupled with atomicabsorption spectrometry and flameless graphite furnace electrothermal atom-isation is described which is essentially 100% quantitative for each of thefour metals studied, has lower blanks and detection limits, and yields betterprecision than previously published techniques. A more precise and accuratedetermination of these metals in seawater at their natural ng/l concentra-tion levels is therefore possible. Samples analysed by this procedure and byconcentration on Chelex 100 showed similar results for cadmium and zinc.Both copper and nickel appeared to be inefficiently removed from seawater byChelex 100. Comparison of the organic extraction results with other pertinentinvestigations showed excellent agreement.


Figure 5.18 is an absorbance versus time plot obtained by Hoenig andWollast [681] for the determination of trace metals in seawater. It shows theabsorbance profiles of the desired elements as a function of the atomisationtemperature. The scale starts with cadmium, for which the absorption signalappears around 400 ◦C, followed by lead (756 ◦C), copper (1000 ◦C), manganese(1200 ◦C), nickel (1300 ◦C), and chromium (140 ◦C).

The time required to completely volatilise the metal increases inversely withthe volatility of the element, and is shown by an enlargement of the absorbancepeak for the more refractory elements. These element profiles are superim-posed with the nonspecific absorption profiles generated by the atomisationof the seawater salts for four preset pyrolysis temperatures. These temperaturesettings cover the temperature range of pyrolysis for the determination of theindicated elements. Figure 5.18 shows that volatile elements caused the mostproblems during analysis because of the incomplete removal of the matrix atthe required low pyrolysis temperatures. Also, the maximum of the nonspe-cific absorption profile does not necessarily coincide with the maxima of theanalyte signals. As in the case of cadmium, the pyrolysis temperature cannotexceed 380 ◦C, but at this pyrolysis temperature the background absorbancecaused by the seawater matrix becomes excessively high (A profile). However,the temperature at which cadmium appeared (corresponding to the top ofthe peak) corresponds to the slope of the background (point Y) at approxi-mately 0.8 units of absorbance (0.8 A), which can be easily compensated bya deuterium corrector.

Figure 5.18. Absorbance signals of test elements compared to background absorbance gen-erated by seawater during atomisation for a pyrolysis performed at 380 ◦C (A), 630 ◦C (B),850 ◦C (C), and 1400 ◦C (D), in the presence of NH4NO3 (4%). Source: [681]


Although the pyrolysis temperature can be raised up to 630 ◦C to reducethe nonspecific absorption (B profile), the determination of lead is a problembecause the appearance temperature of lead corresponds to the maximumabsorbance of the matrix signal (point X). Furthermore, the background ab-sorbance of 1.5 A is at the limit of capability of the deuterium arc corrector.This high background, coupled with the poorer sensitivity of lead (comparedto that of cadmium), limits the analytical capability of the direct determinationof lead in seawater.

The results demonstrate that cadmium can be determined directly; thedirect determination of copper, manganese, and chromium is also possible,but their application is more limited than cadmium. The lead and nickeldetermination proved to be the most difficult, since their determination islimited by their low sensitivity and by the overlap of their absorption profileswith the background absorbance generated by seawater matrix. The directdeterminationof leadandnickel by this technique canbeusedonly for seawatersamples taken in coastal or estuarine zones that are quite polluted.

Furthermore, it is important to emphasise the favourable or unfavourableinfluence of many analytical and instrumental parameters on the quality of theanalysis. It is primarily the state of the graphite tube that can bring about someserious changes regarding magnitude of background during atomisation. Boththe configuration and construction of the atomiser play an important role.Background levels can vary considerably depending on the type of furnaceused. For example, the background is more elevated in older Instrumenta-tion Laboratory atomisers (Models 455 and 555). In contrast, in newer IL655atomisers, the reduction of background absorbance is attributed to the flowprogramming of purge gas during the atomiser cycle. Optimum pyrolysis andatomisation temperatures are so critical to the analysis of seawater that it isnecessary to optimise them for the specific equipment being used.

Kingston et al. [129] have described a method for determining cadmium,cobalt, copper, iron, manganese, nickel, lead, and zinc in seawater usingChelex 100 resin and graphite furnace atomic absorption spectrometry. ThepH of the seawater is adjusted to 5.0 to 5.5 and then passed through a Chelex 100resin column. Alkali and alkaline earth metals are eluted from the resin withammonium acetate and then the trace elements are eluted with two 5 ml por-tions of 2.5 M nitric acid. The difficulties previously encountered with resinswellingandcontractionhavebeenovercome.Bycareful selectionof the instru-mental conditions, it is possible to determine subnanogram levels of these traceelements by graphite furnace atomic absorption spectrometry. The method hasbeen shown to separate quantitatively, with greater than 99% recovery, the ele-ments desired from the alkali and alkaline earth metals, and it has been appliedin the analysis of trace elements in estuarine water from the Chesapeake Bayand seawater from the Gulf of Alaska.

Abollino et al. [690] compared absorptive cathodic stripping voltammetryand graphite furnace atomic absorption spectrometry in the determination of


cadmium, copper, iron, magnesium, nickel, and zinc in seawater. The effectsof UV irradiation and acidification on line preconcentration were studied.

Lead, Magnesium, Vanadium, and Molybdenum

Tominaga et al. [682,683] studied the effect of ascorbic acid on the response ofthese metals in seawater obtained by graphite-furnace atomic absorption spec-trometry from standpoint of variation of peak times and the sensitivity. Matrixinterferences from seawater in the determination of lead, magnesium, vana-dium, and molybdenum were suppressed by addition of 10% (w/v) ascorbicacid solution to the sample in the furnace. Matrix effects on the determinationof cobalt and copper could not be removed in this way. These workers proposea direct method for the determination of lead, manganese, vanadium, andmolybdenum in seawater.

Copper, Iron, Manganese, Cobalt, Nickel, and Vanadium

Segar and Gonzalez [431] carried out a direct determination of these elementsin seawater using a graphite atomiser and a deuterium background connec-tor. Sea salts are volatilised at a lower temperature than is required for thevolatilisation of the above elements.

Nickel, Copper, Molybdenum, and Manganese

Hayase et al. [684] first extracted the seawater sample with chloroform toremove dissolved organic matter prior to analysis of the aqueous phase bygraphite furnace atomic absorption spectrometry. Seawater samples at pH 3and at pH 8 were extracted with chloroform, evaporated to dryness, and theresidue treated with nitric acid. Acid solutions were subjected to metal analysesby graphite furnace atomic absorption spectrometry.

Mercury, Lead, and Cadmium

Tikhomirova et al. [685]developedaprocedure for simultaneous concentrationof mercury, lead, and cadmium from seawater by coprecipitation with coppersulfide. The isolation yield is 99% for mercury and lead, and 89% for cadmium.Mercury isdeterminedbyflamelessatomicabsorptionspectrophotometry, andlead and cadmium by flame atomic absorption spectrophotometry.

Arsenic, Bismuth, Indium, Lead, Antimony, Selenium, Tin, Tellurium, and Thallium

The application of palladium and magnesium nitrate matrix modifier forgraphite furnace atomic absorption spectrometry has been discussed in de-tail [686]. The work has shown that a mixture of palladium and magnesium


nitrates is a powerful matrix modifier for nine elements of Group IIIA throughGroup VIA of the periodic table. Preliminary results show good promise thatthismodifiermightbeused forevenmoreelements, including thoseofGroupIBand Group IIB. This would mean that the palladium and magnesium nitratesmodifier could possibly replace almost all the other matrix modifiers rec-ommended up to now, except perhaps for the magnesium nitrate modifierproposed for several transition elements [687]. This would certainly simplifygraphite furnace AAS in routine applications.

The palladium and magnesium nitrates modifier makes it possible to applythermal pretreatment temperatures of at least 900–1000 ◦C for all investigatedelements. For most elements this modifier supports substantially higher py-rolysis temperatures than did the matrix modifier recommended previouslyby Perkin-Elmer [688]. These higher pyrolysis temperatures allow for effectivecharring of biological matrices and removal of most inorganic concomitantsprior to analyte element volatilisation.

The palladium and magnesium nitrates modifier has a substantial equal-ising effect on the atomisation temperature of the nine elements investigated.The optimum atomisation temperature for all but one element (thallium) isbetween 1900 and 2100 ◦C. This means that all elements can be determined ata “compromise” atomisation temperature of 2100 ◦C with a minimum sacri-fice in sensitivity. Such uniform conditions for as many elements as possibleare of vital importance if simultaneous multielement furnace techniques areenvisaged. Moreover, in conventional graphite furnace AAS, uniform condi-tions for a number of elements can greatly facilitate and simplify daily routineanalysis.

The mechanism of stabilisation of the palladium and magnesium nitratesmodifier was not investigated [749]. It is known, however, that palladiumnitrate decomposes via the oxide to the metal at 870 ◦C, which melts at 1552 ◦C.The appearance temperature for palladium in a graphite furnace is around1250 ◦C. As most of the investigated elements are stabilised to temperaturesaround 1200 ◦C, it can be assumed that the modifier acts by imbedding theanalyte into the palladium matrix, or even by forming a kind of “alloy” withthe analyte.

Platinum and Iridium

Hodge et al. [689] have described a method for the determination of platinumand iridium at picogram levels in marine samples, based upon an isolationof anionic forms of these elements using appropriate resins, with subsequentpurification by uptake on a single ion-exchange bead. All steps are followed byradiotracers, and yields vary between 35 and 90%. Graphite furnace AAS wasemployed as the determinative step.


5.74.6Zeeman Graphite Furnace Atomic Absorption Spectrometry

The widespread use of graphite furnace systems has greatly expanded therequirements for accurate background correction in atomic absorption mea-surements. Correction for background absorption is most commonly achievedusing continuum sources. While adequate in many cases, the continuum sourcetechnique has several inherent limitations. The intensity of the continuumsources is not always adequate; inaccurate correction is possible if the back-ground is structured; plus it is necessary to maintain correct optical alignmentbetween the source and continuum lamps. The combined effect of these lim-itations means that it is not always possible to obtain accurate correction forapplications with high background levels.

The need for improved background correction performance has generatedconsiderable interest in applying the Zeeman effect, where the atomic spectralline is split into several polarised components by the application of a magneticfield. With a Zeeman effect instrument background correction is performedat, or very close to, the analyte wavelength without the need for auxiliary lightsources. An additional benefit is that double-beam operation is achieved witha very simple optical system.

In 1971 Hedeishi and McLaughlin [457] first reported the application of theZeeman effect for the determination of mercury. Numerous workers have sinceinvestigated the technique, utilising systems in which the magnetic fields wereapplied directly to the light source [692–700] or to the atom source [701–704].There are several possible design approaches for a Zeeman effect atomic ab-sorption spectrophotometer. The magnetic field may be fixed or modulated,the field may be aligned in a direction transverse (perpendicular) or longitu-dinal (parallel) to the optical path, and the field may be applied to the lightsource or the atom source.

Pernandez et al. [705], of Perkin-Elmer Limited, reported results obtained ininvestigating several of these possible design approaches. Background correc-tion performance will probably not be significantly influenced by the positionor type of magnet used. However, this is not the case regarding sensitivity andanalytical range, where the design employed will have a significant impact onperformance. These workers developed a Zeeman effect instrument capable ofproviding improved background correction performance with minimal sacri-fice in analytical sensitivity or working range, and this was incorporated intothe Model 5000 instrument. This design utilises a modulated transverse fieldapplied to the graphite tube. Comparisons of analytical sensitivity and work-ing range versus standard atomic absorption performance with the system arereported for the determination of manganese in seawater [452].

Further improvements in the technique [706] includes the use of a L’vovplatform to achieve a temperature that is constant in time, and improved py-rolytically coated graphite tubes. To achieve improved performance requires


a fast spectrophotometer, rapid heating of the furnace, integration of the ab-sorbance signals, and usually an appropriate matrix modifier. All of theseaspects of the analytical system must be carefully integrated with an under-standing of the role played by the system. These workers give guidelines foroptimising each part of the system.

In addition to manganese, discussed below [452], the Zeeman techniquehas been used for the determination of other elements [707] in seawater.

Heavy Metals

De Kersabiec et al. [708] have described a Zeeman method for the determina-tion of copper, lead, cadmium, cobalt, nickel, and strontium in brines and inthe soil water adjacent to the Red Sea.

Manganese

To determine manganese [452] several factors had to be controlled carefullyto obtain reliable results against simple standards that were independent ofsalinity and variations in matrix composition. Use of the L’vov platform andintegration of the absorbance signal reduced the sensitivity to matrix com-position. Pyrolytically coated graphite reduced variations that depend uponthe life of the tubes. The tubes appeared to fail by intercalation of the Na orNaCl matrix. The char temperature must not vary outside the range 1100–1300 ◦C (see below). Zeeman background correction permitted use of largerseawater samples. The detection limit of the procedure using 20 µl samples was0.1 µg/l (2 pg) manganese. By use of the Zeeman background corrector, lessthan 0.02 µg/l manganese was detected in seawater using a 75 µl sample.

Chromium, Nickel, Manganese, Cadmium, Arsenic, and Molybdenum

Grobenski et al. [709] have reviewed methodology for the determination ofthese elements in seawater. Zeeman-effect background correction using an ACmagnet around the graphite furnace corrects for nonspecific attenuation up to2.0 absorbance and corrects for structured background.

Grobenski et al. [709] point out that in analysing different seawater sam-ples under standard temperature platform furnace conditions, one does notalways have to deal with such a high background. On the other hand, “over-compensation” using a continuum source background corrector for modestbackground is usually ascribed to the presence of structured background. Ina few cases it is very difficult to distinguish between fast background andstructured background.

The following additional, even more important, requirements are put onbackground correction, namely: there should be no analytical sensitivity loss


associated with correction; accuracy of correction must be very good; and cor-rection of fast background signals should be possible. Normal concentrationsfor a number of elements in seawater is near or only slightly above the detectionlimits.

Maximum power heating, the L’vov platform, gas stop, the smallest possibletemperature step between thermal pretreatment and atomisation, peak areaintegration, and matrix modification have been applied in order to eliminateor at least reduce interferences in graphite furnace AAS. With Zeeman effectbackground correction, much better correction is achieved, making methoddevelopment and trace metal determinations in samples containing high saltconcentrations much simpler or even possible at all.

Molybdenum is an element for which platform atomisation does not offeran advantage. Just the opposite is the case; sensitivity is very poor and memoryeffects are very strong. The Zeeman detection limit for wall atomisation ina pyrocoated graphite tube using 100 µl of reference solution is 0.03 µl (forboth peak height and peak area evaluation) [709].

The molybdenum concentration in the reference sample is rather high anda direct determination using the “cookbook” conditions [710] is very straight-forward. There is no difference between peak area and peak height evaluation.In spite of 1800 ◦C for thermal pretreatment, a small background absorptionsignal is present.

The experimental value of 11.7 µg/l obtained by this technique is in excellentagreement with the reference value of 11.5 µg/l, and the detection limit was0.04 µg/l.

Cabon and Le Bihan [711] studied the effects of transverse heated AAS andlongitudinal Zeeman effect background correction in sub µg/l determinationof chromium, copper, and manganese in seawater samples.


Arsenic, Antimony, Bismuth, Selenium, Tellurium, Tin, Lead, and Germanium

A number of elements in the fourth, fifth, and sixth group of the periodictable (Ge, Sn, Pb, As, Sb, Se, Te) form hydrides upon reduction with sodiumborohydride, which are stable enough to be of use for chemical analysis. Ofthese elements, Andreae [712] has investigated in detail arsenic, antimony,germanium, and tin. The inorganic and organometallic hydrides are separatedby a type of temperature-programmed gas-chromatography. In most cases itis optimal to combine the functions of the cold trap and the chromatographiccolumn in one device. The hydrides are quantified by a variety of detectionsystems, which take into account the specific analytical chemical propertiesof the elements under investigation. For arsenic, excellent detection limits(≈ 40 pg) can be obtained with a quartz tube cuvette burner, which is po-


sitioned in the beam of an atomic absorption spectrophotometer. For someof the methylarsines, similar sensitivity is available with an electron capturedetector. The quartz-burner/AAS system has a detection limit of 90 pg for tin;for this element much lower limits (≈ 10 pg) are possible with a flame pho-tometric detection system, which uses the extremely intense emission of theSnH molecule at 609.5 nm. The formation of GeO at the temperatures of thequartz tube furnace makes this device quite insensitive for the determinationof germanium. Excellent detection limits (≈ 140 pg) can be reached for thiselement by the combination of the hydride generation system with a modifiedgraphite furnace/AAS.

The application of these techniques has led to the discovery of a number oforganometallic species of arsenic, tin, and antimony in the marine environ-ment. Germanium has not been observed to form organometallic compoundsin nature. Some aspects of the geochemical cycles of these elements which havebeen elucidated by the use of these methods are discussed.

Braman et al. [713] suggested the use of sodium borohydride (NaBH4) asa reducing agent to replace the metallic zinc used in the classical Marsh test,which is awkward to handle and often contains large blanks of the elementsof interest. Sodium borohydride is now used almost exclusively in the variousmodifications of the hydride method.

Many of the recent methods make use of the condensation of the hydridesin a cold trap at liquid nitrogen temperature. Braman and Foreback [714]pioneered the use of a packed cold trap to serve both as a substrate tocollect the hydrides at liquid nitrogen temperature, and to separate arsineand the methylarsines chromatographically by controlled heating of the trap.In the same paper, they described the differentiation between arsenic (III)and arsenic (V) by a prereduction step and by control of the pH at whichthe reduction takes place. A variety of highly sensitive detectors are used,many of which are element-selective. Most of the detectors commonly usedfor gas chromatography have been applied to the detection of the hydrides,among them thermal conductivity, flame ionisation, and electron capture de-tector. A molecular emission detector has been used for tin. Atomic emissionspectrometric detectors based on DC discharges [714, 715] and microwave-induced plasmas were applied to the speciation of arsenic in environmen-tal samples. The currently most popular detection system is AAS in one ofits numerous variants. The hydrides were at first introduced into a normalAA flame, but it was soon recognised that better detection limits could beachieved with enclosed atom reservoirs and with very small flames or withflameless systems. A number of heated quartz furnace devices without inter-nal flames are now on the commercial market. The lowest detection limitswere achieved by cold-trapping of the hydrides and subsequent introductioninto either a quartz cuvette furnace [716] or into a commercial graphite fur-nace [715]. Andreae [712] discusses the methodology of the determinationof arsenic, antimony, germanium, and tin with these systems, and its appli-


Figure 5.19. Schematic design of the NaBH4-reduction/flame photometric detection systemfor the determination of tin species in natural waters. Source: Author’s own files

cation to the investigation of the marine and estuarine chemistry of theseelements.

The determination of the hydride element species consists of five steps:

1. Reduction of the element species to the hydrides2. Removal of interferent volatiles from the gas stream3. Cold-trapping of the hydrides4. Separation of the substituted and unsubstituted hydrides from each other

and from interfering compounds5. Quantitative detection of the hydrides

Andreae discusses each of these steps in detail [712]. A typical instrumentalconfiguration to accomplish these steps is shown in Fig. 5.19 for the borohy-dride reduction/flamephotometricdetectionsystemfor tin speciationanalysis.

Reduction of the Element Species to the Hydrides

Most of the hydride elements occur in a number of different species. The opti-mum reduction conditions vary from element to element, and among differentspecies of the same element, e.g., antimony (III), antimony (V), methylsti-bonic acid ([(CH3)SbO(OH)2]) and dimethylstibinic acid [(CH3)2SbO(OH)].The conditions under which the element species are being reduced have beenoptimised as shown in Table 5.5.

With the exception of antimony (V), which requires the presence of iodidefor its reduction, all species can be reduced in an acid medium at a pH of 1–2.However, the reduction of some species, including antimony (III), arsenic (III),and all tin species, will also proceed at higher pH, where arsenic (V) andantimony (V) are not converted to their hydrides. This effect permits theselective determination of the various oxidation states of these elements [714,716]. In the case of tin, reduction can be achieved at the pH of the Tris-HC1


buffer (∼6–7), but due to the tin contamination in commercially available Tris-HC1, Andreae [712] prefers to perform the reaction in a medium containinga small amount of nitric acid. This addition results in a solution pH of about8 after the injection of the NaBH4; without it, the pH would rise above 10 andthe reduction to the stannanes would be inhibited. Nitric acid is used, as itis available with a tin blank below the limit of detection (most HC1 containsdetectable tin blanks). The acid is added to the sample immediately after it hasbeen taken; it then serves both to stabilise the solution and to control the pHof the analytical reaction. Andreae [712] was not able to differentiate betweenSn (II) and Sn (IV); both species are reduced with the same yield under hisoperating conditions.

In contrast to the findings of Foreback [720] and Tompkins [715], An-dreae [712] was not able to reduce antimony (V) quantitatively at pH 1.5–2without the addition of potassium iodide. A concentration of at least 0.15 M KIin the final solution at a pH less than 1.0 was necessary to achieve com-plete reduction [716]. This is in agreement with the work of Fleming andIde [717] who suggested an addition of approximately 0.2 M KI per litre toensure the reduction of Sb (V). Under the conditions used by Foreback [614],Andreae [712] finds only partial reduction of both Sb (III) and Sb (V) (about30% for both species).

Table 5.5. Reaction conditions for the reduction of various ‘hydride element’ species to thecorresponding hydrides

Species1 pKa pH Composition of NaBH4reaction medium

As(III) 9.2 6–7 0.05 M TRIS-HCl 1As(V) 2.3 ∼1 0.12 M HCl 3MMAA 3.6DMAA 6.2Sn(II) 9.5 2–8 0.01 M HNO3 1Sn(IV) ∼10MexSn 11.71

Ge(IV) 9.3 ∼6 0.095 M TRIS-HCl 3Sb(III) 11.0 ∼6 0.095 M TRIS-HCl 2Sb(V) 2.7 ∼1 0.18 M HCl, 0.15 M KI 3MMSA – 1.5–2.0 0.06 M HCl 2DMSA –

MMAA: monomethylarsonic acid [(CH3)AsO(OH)2]DMAA: dimethylarsinic acid [(CH3)2AsO(OH)]MexSn: MeSn3+, Me2Sn2+, Me3Sn+

MMSA: monomethylstibonic acid [(CH3)SbO(OH)]DMSA: dimethylstibonic acid [(CH3)2SbO(OH)]1Data available only for Me2Sn(OH)2From Author’s own files


Germanium can be reduced through a wide range of pH [716]. The optimumpH is in the near-neutral range, as the efficiency of germanium reduction de-creases at lower pH, probably due to the competitive acid-catalysed hydrolysisof the borohydride ion. At a pH above 8, the yield also decreases, both dueto the decrease in reducing power of borohydride with increasing pH, and tothe lack of hydrogen evolution at high pH (the hydrogen gas evolving in thesolution helps to strip out the hydrides more efficiently than the relatively largehelium bubbles passing through the solution).

Removal of Interferent Volatiles from the Gas Stream

Depending on the detector used, some volatile compounds which are formedor released during the hydride generation step may interfere with the detec-tion of the hydrides of interest. Most prominent among them are water, carbondioxide, and in the case of anoxic water samples, hydrogen sulfide. The atomicabsorption detector is insensitive to these compounds, so no precautions needbe taken when this detector is used. It has been found convenient in someapplications, however, to remove most of the water before it enters the cold-trap/column, which serves to condense and separate the hydrides. This canbe accomplished by passing the gas stream through a larger cold trap cooledby a dry ice/alcohol mixture or by an immersion-cooling system [716]. Thismethod was also used with water-sensitive detectors, e.g., the electron-capturedetector for methylarsines [716], or with plasma discharge detectors (e.g.,Crecelius [718]). Carbon dioxide produces an interfering peak on the plasmadischarge detectors and the flame photometric detector for tin. If the separa-tion by the column used is adequate, no additional precautions are necessaryto remove carbon dioxide interference. Otherwise, a small tube filled withgranulated sodium hydroxide can be included in the gas stream to absorb car-bon dioxide. Samples of marine anoxic water often contain large amounts ofhydrogen sulfide. This causes a significant interference in a number of differentdetectors. It can be removed by passing the gas through a tube filled with leadacetate [718].

Cold Trapping of the Hydrides

Only when the very contamination-sensitive electron-capture detector is usedis it necessary to provide separate gas streams, one for the reaction and strip-ping part of the system, the other for the carrier gas stream of the column anddetector. Otherwise, the same gas stream can be used to strip the hydridesfrom solution and carry them into the detector, which greatly simplifies theapparatus. This is of considerable significance, as each additional surface andjoint in the apparatus increases the possibility of irreversible adsorption of thesensitive hydrides, and thus is a potential contributor to analytical error. The


Figure 5.20. a Chromatogram of stibine, methylstibine, and dimethylstibine as separated bythe OV-3 trap/column with the quartz furnace/AA detector. The stibines were prepared bythe NaBH4 reduction of 2 ng Sb each as Sb (III), methylstibonic, and dimethylstibonic acid.b Analysis of a sample of seawater (100 ml) from the open Gulf of Mexico (Station SN4-3-1,12 March 1981)

cold trap then serves both to collect the hydrides from the reaction gas streamand chromatographically separate them as it is heated up. Initially, columnpackings of glass beads [614] or glass wool [716] were used. These packingsproduce poor separation of the methylated species from one another, and badlytailing peaks. Andreae [712] therefore used a standard gas chromatographicpacking (15% OV-3 on Chromosorb W/AWDMCS, 60–80 mesh) in U-tubes forthe separation of the inorganic and alkyl species of arsenic, antimony, and tin.This packing is quite insensitive to water and produces sharp, well-separatedpeaks, as demonstrated in Fig. 5.20 for stibine, methylstibine, and dimethyl-stibine in both standard and seawater samples. The retention times can beregulated by winding a heating wire around the outside of the U-tube andcontrolling the current supplied to this heating coil.

Quantitative Detection of the Hydrides

Andreae [712] used four different detectors in his investigations: the electroncapture detector (for the methylarsines), the quartz cuvette atomic absorp-tion detector (for arsenic and antimony species), the graphite furnace atomic


absorption detector (for germanium and tin species), and the flame photomet-ric detector (for tin species). Their performance in the borohydride analysissystem was evaluated.

The electron-capture detector was originally found to be a sensitive detectorfor the methylarsines [716]. After improvements of the atomic absorptiondetectors had been made (especially concerning adsorptive losses and peakshapes of the methylarsines), it was found that this detector could be used toreplace the electron-capture detector, which because of its lack of specificityand its sensitivity to contamination and changes in operating conditions wasvery inconvenient to work with.

The most versatile system is the combination of hydride generation withAAS. Here, the objective is to introduce the hydrides into an atom reservoiraligned in the beam of the instrument and to dissociate them to producea population of the atoms of interest. This can be achieved either in a fuel-richhydrogen/air flame in a quartz tube (cuvette), as described by Andreae [712],or in a standard graphite furnace by electrothermal atomisation [716]. Thequartz cuvette has higher sensitivity than the graphite furnace for arsenicand antimony; it is therefore preferred for the determination of these twoelements. When organotin compounds are analysed using the quartz cuvettesystem, spurious peaks are sometimes seen eluting with the methyltins. Theorigin of these peaks is not clear. This interference can be avoided by using thegraphite furnace system. Here, the hydrides are introduced with the carriergas stream, to which some argon has been added, into the internal purge inletof the graphite furnace. They have to pass through the graphite tube and leavethrough the internal purge outlet. The heating cycle of the furnace is timedso it reaches the required atomisation temperature shortly before the arrivalof the unsubstituted hydride, and is held at temperature until the last alkyl-substituted hydride has eluted. With this system, probably due to the higheroperating temperature, no spurious tin peaks are present.

The graphite furnace system was originally developed by Andreae [712]when he found that the quartz cuvette gave only very poor sensitivity for ger-manium. This was attributed to the formation of GeO, a very stable diatomicspecies, at the relatively low temperatures of the quartz cuvette. At the highertemperatures available with the graphite furnace (2600 ◦C for the determina-tion of Ge), a sensitivity could be obtained for germanium comparable to thatof the other hydride elements.

Nakashima et al. [719] detail a procedure for preliminary concentration of16 elements from coastal waters and deep seawater, based on their reductiveprecipitation by sodium tetrahydroborate, prior to determination by graphite-furnace AAS. Results obtained on two reference materials are tabulated. Thiswasa simple, rapid, andaccurate technique fordeterminationof awide rangeoftrace elements, including hydride-forming elements such as arsenic, selenium,tin, bismuth, antimony, and tellurium. The advantages of this procedure overother methods are indicated.



The DC plasma was introduced as an excitation source for atomic emis-sion spectrometry by Margoshes and Scribner [721] and Korolev and Vain-shtein [722]. Modified designs have been characterised by a number of otherauthors [614, 719–729]. Commercial equipment is now available from severalmanufacturers. The principle of the plasma torch arrangement used in theseinstruments is illustrated in Fig. 5.21 [730].

Winge et al. [730] have investigated the determination of twenty or moretrace elements in saline waters by the inductively coupled plasma technique.They give details of experimental procedures, detection limits, and precisionand accuracy data. The technique when applied directly to the sample is notsufficiently sensitive for the determination of many of the elements at the lowconcentrations at which they occur in seawater, and for these samples precon-centration techniques are required. However, it has the advantages of beingamenable to automation and capable of analyzing several elements simultane-ously.

Figure 5.21. Torch and sam-ple aerosol generation sys-tem (QVAC 127 system).Source: [510]


Heavy Metals

The application of the Spectroscan DC plasma emission spectrometer con-firmed that for the determination of cadmium, chromium, copper, lead, nickel,and zinc in seawater the method was not sufficiently sensitive, as its detectionlimits just approach the levels found in seawater [731]. High concentrationsof calcium and magnesium increased both the background and elemental lineemission intensities.

The extension of inductively coupled plasma (ICP) atomic emission spec-trometry to seawater analysis has been slow for two major reasons. The first isthat the concentrations of almost all trace metals of interest are 1 µg/l or less,below detection limits attainable with conventional pneumatic nebulisation.The second is that the seawater matrix, with some 3.5% dissolved solids, is notcompatible with most of the sample introduction systems used with ICP. Thusdirect multielemental trace analysis of seawater by ICP–AES is impractical,at least with pneumatic nebulisation. In view of this, a number of alternativestrategies can be considered:

1. Preconcentration and removal of the metals of interest from the seawatermatrix prior to ICP analysis

2. Use of ultrasonic nebulisation with aerosol desolvation3. Combination of the foregoing two strategies

Owing to inadequate detection limits by direct analysis, various workersexamined preconcentration procedures, including dithiocarbamate precon-centration [447,732–734], ion exchange preconcentration [735–737], chelationsolvent extraction [736], coprecipitation [738], and preconcentration in silica-immobilised 8-hydroxyquinoline [129].

Berman et al. [735] have shown that if a seawater sample is subjected to20-fold preconcentration by one of the above techniques, then reliable analysiscan be performed by ICP–AES (i.e., concentration of the element in seawateris more than five times the detection limit of the method) for iron, manganese,zinc, copper, and nickel. Lead, cobalt, cadmium, chromium, and arsenic arebelow the detection limit and cannot be determined reliably by ICP–AES. Theselatter elements would need at least a hundredfold preconcentration before theycould be reliably determined.

Berman et al. [735] and McLaren [738] attempted to determine the foregoingnine elements in seawater by a combination of ion exchange preconcentrationon Chelex 100 [129,736–738], and ICP–AES using ultrasonic nebulisation. Pre-concentration factors of between 25 and 100 were obtained by this technique.

Table 5.6 compares the ICP–AES results with data generated for the samesampleby twoother independentmethods– isotopedilutionspark sourcemassspectrometry (IDSSMS), andgraphite furnaceatomicabsorptionspectrometry(GFAAS). The IDSSMS method also uses 25-fold preconcentration of the metalsand matrix separation using the ion exchange procedure, following isotope


Table 5.6. Analysis of Sandy Cove seawater

Element ICP-AES GFAAS IDSSMS

Mn 1.5 ± 0.1 1.4 ± 0.2 1.8 ± 0.2∗Fe 1.5 ± 0.6 1.5 ± 0.1 1.4 ± 0.1Zn 1.5 ± 0.4 1.9 ± 0.2 1.6 ± 0.1Cu 0.7 ± 0.2 0.6 ± 0.2 0.7 ± 0.1Ni 0.4 ± 0.1 0.33 ± 0.08 0.37 ± 0.02Pb – 0.22 ± 0.04 0.35 ± 0.03Cd – 0.24 ± 0.04 0.28 ± 0.02Cr – – 0.34 ± 0.06Co – – 0.02∗

Precision expressed as 95% confidence intervals∗ Spark source mass spectrometry, internal standard methodFrom [735]

addition. The atomic absorption determinations were preceded by an methyliosbutyl ketone extraction [736]. In general agreement is very good, but onediscrepancy merits comment. The spark source mass spectrometry result formanganese is not as reliable as the other data by this method. Since mangneseis monoisotopic, a less accurate internal standardisation method of calibrationhas been used. The ICP–AES result for manganese is in close agreement withthe GFAAS result.

Sturgeon et al. [736] compared five different analytical methods in a studyof trace metal contents of coastal seawater. Analysis for cadmium, zinc, lead,iron, manganese, copper, nickel, cobalt, and chromium was carried out usingisotope dilution spark source mass spectrometry (IDSSMS), graphite furnaceatomic absorption spectrometry (GFAAS), and inductively coupled plasmaemission spectrometry (ICPES) following trace metal separation preconcen-tration (using ion exchange and chelation solvent extraction) and direct anal-ysis (by GFAAS).

Table 5.7 gives results obtained on a sample of seawater. Overall, there isgood agreement in elemental analysis obtained by the various methods.

Although ICP–AES is a multielement technique, its inferior detection lim-its relative to GFAAS would necessitate the processing of large volumes ofseawater, improvements in the preconcentration procedures in use thus far,or new, alternative preconcentration procedures such as carrier precipitation(see below).

Hiraide et al. [737] developed a multielement preconcentration techniquefor chromium (III), manganese (II), cobalt, nickel, copper (II), cadmium, andlead in artificial seawater using coprecipitation and flotation with indium hy-droxide followed by ICP–AES. The metals are simultaneously coprecipitatedwith indium hydroxide adjusted to pH 9.5, with sodium hydroxide, ethano-lic solutions of sodium oleate and dodecyl sulfate added, and then floated to


Table 5.7. Analysis of seawater sample

Element Concentration (µg/l)

GFAAS ICPES ion IDSSMS ionexchange exchange

Direct Chelation – extraction

Fe 1.6 ± 0.2∗ 1.5 ± 0.1 1.5 ± 0.6 1.4 ± 0.1Mn 1.6 ± 0.1 1.4 ± 0.2 1.5 ± 0.1 NDCd 0.20 ± 0.04 0.24 ± 0.04 ND 0.28 ± 0.02Zn 1.7 ± 0.2 1.9 ± 0.2 1.5 ± 0.4 1.6 ± 0.1Cu ND 0.6 ± 0.2 0.7 ± 0.2 0.7 ± 0.1Ni ND 0.33 ± 0.08 0.4 ± 0.1 0.37 ± 0.02Pb ND 0.22 ± 0.04 ND 0.35 ± 0.03Co ND 0.018 ± 0.008† ND 0.020 ± 0.003‡

∗: Precision expressed as standard deviation†:Pre-concentrated 100-fold by Chelex 100 ion exchange‡: Spark source mass spectrometry, internal standard methodFrom [736]

the solution surface by a stream of nitrogen bubbles. Cadmium may be com-pletely recovered without the coprecipitation of magnesium. Concentrationsof the heavy metals chromium (III), manganese (II), cobalt, nickel, copper (II),cadmium, and lead in 1200 ml of artificial seawater were increased 240-fold,while those of sodium and potassium were reduced to 2–5%, and those ofmagnesium, calcium, and strontium were reduced to 50%.

Down to 1 µg/l of the aforementioned heavy metals can be determined bythis procedure. However, it is emphasised that real seawater samples were notincluded in this study.

Heavy Metals and Vanadium

Sugimae [447] developed a method for heavy metals in which they are chelatedwith diethyldithiocarbamic acid, the chelates are extracted with chloroform,and each chelate decomposed prior to determination. When 1 litre watersamples are used, the lowest determinable concentrations are: Mn (0.063 µg/l),Zn (0.13 µg/l), Cd (0.25 µg/l), Fe (0.25 µg/l), V (0.38 µg/l), Ni (0.5 µg/l), Cu(0.5 µg/l), and Pb (2.5 µg/l). Above these levels the relative standard deviationsare better than 12% for the complete procedure.

Heavy Metals, Molybdenum, and Vanadium

Mujazaki et al. [733] found that di-isobutyl ketone is an excellent solvent for theextraction of the 2,4-pyrrolidone dithiocarbamate chelates of these elementsfrom seawater.


Unlike halogenated solvents, it does not produce noxious substances inthe inductively coupled plasma, has a very low aqueous solubility, and yieldshundredfold concentration in one step. Detection limits ranged from 0.02 µg/l(cadmium) to 0.6 µg/l (lead). The results indicate that the proposed procedureshould be useful for the precise determination of metals in oceanic water,although a higher sensitivity would be necessary for lead and cadmium.

A comparison was carried out on the results obtained using ICP–AES andAAS for eight elements in coastal Pacific Ocean water. The results for cadmium,lead, copper, iron, zinc, and nickel are in good agreement. For iron, the dataobtained by the solvent extraction ICP method are also in good agreementwith those determined directly by ICP–AES. In most of the results the relativestandard deviations were 4% for all elements except cadmium and lead, whichhad relative standard deviations of about 20% owing to the low concentrationsdetermined.

Arsenic, Antimony, and Selenium

De Oliviera et al. [739] have described a technique for determining theseelements based on the hydride generation technique. Detection limits are1 µg/l for arsenic and antimony, and 0.5 µg/l for selenium.

Bismuth, Cadmium, Copper, Cobalt, Indium, Nickel, Lead, Thallium, and Zinc

Berndt et al. [740] have shown that traces of bismuth, cadmium, copper, cobalt,indium, nickel, lead, thallium, and zinc could be separated from samples ofseawater, mineral water, and drinking water by complexation with the ammo-nium salt of pyrrolidine-1-dithiocarboxylic acid, followed by filtration througha filter covered with a layer of active carbon. Sample volumes could range from100 ml to 10 litres. The elements were dissolved in nitric acid and then deter-mined by atomic absorption or inductively coupled plasma optical emissionspectrometry.


Although the use of inductively coupled plasma mass spectrometry (ICP-MS)is rapidly expanding, because of the many attractive features of this technique,its application to the analysis of saline waters remains limited. This is largelydue to the low tolerance of the technique to dissolved solids, with the highestrecommended level being 0.2%, if a solution is to be continuously nebulisedwithout inducing undue instrumental drift caused by solid deposition on theorifice. Another restriction comes from effects of concomitant elements thatare non-spectroscopic interferences, often resulting in a suppression of ana-lyte signals. Thus, the analysis of seawaters requires a preliminary treatment


in order to reduce their salt content prior to analysis by ICP-MS. This canbe accomplished by, for example, preconcentration on silica-immobilised hy-droxyquinoline, a technique that allows the concentration of a number of tracemetals while separating them from the univalent major ions and, to someextent, the divalent ions such as calcium and magnesium. This technique hasbeen successfully applied to the analysis of the coastal seawater reference mate-rial and the open ocean water reference material NASS-2. It presents, however,the disadvantages of being time-consuming and requiring large volumes ofsample.

Flow injection analysis can be used to speed up the preconcentration processand reduce sample consumption.

Bloxham et al. [842] have reviewed the application of ICP-MS to the deter-mination of trace metals in seawater.

Heavy Metals

Gee and Bruland [953] used 61Ni, 65Cu, and 68Zn in waters collected in SanFrancisco Bay to trace the kinetics of nickel, copper, and zinc exchange betweendissolved and particulate phases. The technique involved an organic ligandsequential extraction followed by analysis with high-resolution ICP-MS.

A similar approach has been used [954] to examine the partitioning of 68Zn,111Cd, and 207Pb between seawater and various organic reservoirs in the mussel(Mytilus galloprovincialis).

Warnken et al. [956] have reported an online preconcentration – ultrasonicnebulisation – ICP-MS method that achieved detection limits of 0.26, 0.86, 1.5,10, and 0.44 ng/l for manganese, nickel, copper, zinc, and lead in seawater. Thisonline preconcentration method compares favourably to the state of-the-artoff-line methods.

Liu et al. [955] developed an electrothermal vaporisation istope dilution –ICP-MS method for determining cadmium, mercury, and lead in seawater at2, 5, and 1 ng/l detection limits, respectively.

Zinc, Manganese, Cobalt, Copper, Chromium, Nickel, Iron, Cadmium, Lead,and Mercury

Chong et al. [742] have described a multielement analysis of multicompo-nent metallic electrode deposits, based on scanning electron microscopy withenergy dispersive X-ray fluorescence detection, followed by dissolution andICP-MS detection. Application of the method is described for determinationof trace elements in seawater, including the above elements. These elementsare simultaneously electrodeposited onto a niobium-wire working electrodeat –1.40 V relative to an Ag/AgCl reference electrode, and subjected to energydispersive X-ray fluorescence spectroscopy analysis. Internal standardisation


is practical for quantitative calibration at the 1 ppm analyte concentration levelin an analyte–internal standard concentration ratio range of 0.02–50. Detec-tion limits for energy dispersive X-ray fluorescence spectroscopy range from1.9 µg/l for iron to 50 µg/l for cadmium. The deposit is dissolved for subsequentICP-MS determination. Significant reduction in ICP-MS matrix interferencesby sodium, calcium, magnesium, potassium, and chloride ions is achievedby deposition at potentials more positive than their very negative reductionpotentials. Measurement of elemental isotope ratios is achieved with 0–8%relative error. ICP-MS detection limits for all elements except zinc and iron aresuperior to those of energy dispersive X-ray fluorescence spectroscopy. Man-ganese, nickel, cadmium, lead, and mercury can easily be determined over therange 13–86 parts per trillion with ICP-MS.

Copper, Cobalt, Manganese, Nickel, Vanadium, Molybdenum, Cadmium, Lead,and Uranium

Beauchemin and Berman [741] used ICP-MS with online preconcentration,employing a miniature column packed with 8-hydroxy quinoline to determinethese elements in open ocean water. This technique improved detection limitsof several elements by a factor of 5–7 compared to ICP-MS alone. The onlinepreconcentration system was first assessed by using the method of standardadditions to determine manganese, cobalt, nickel, copper, lead, and uranium inthe riverine water, SLRS-1, whose salt content was low enough to allow moni-toring of both the preconcentration and the solution processes. Results in goodagreement with the certified values were obtained for all but nickel because ofa spectral interference by calcium oxide from co-eluted calcium. The systemwas successfully applied to the determination of manganese, molybdenum,cadmium, and uranium in the reference open ocean water, NASS-2, by usingan isotope dilution technique and the method of standard additions.

Chapple andByrne [743] appliedanelectrothermal vaporisation inductivelycoupled plasma technique to the determination of copper, cobalt, manganese,nickel, and vanadium in seawater in amounts down to 3–140 ppt.

Heavy Metals, Beryllium, Bismuth, Gallium, Mercury, and Indium

Apoly(acrylaminophosphamic-dithiocarbamate) chelatingfibrehasbeenusedto preconcentratrate several trace metals in seawater by a factor of 200 [957].The elements included beryllium, bimuth, cobalt, gallium, silver, lead, cad-mium, copper, manganese, and indium. ICP-MS was used for detection.

Heavy Metals, Barium

Esser and Volpe [958, 959] used a shipboard single-collector quadrupole ICP-MS to survey surface barium and toxic metal levels in seawater.


Heavy Metals, Vanadium

Field et al. [747] used ICP high-resolution mass spectrometry to determinevanadium, chromium, manganese, iron, cobalt, nickel, copper, zinc, cadmium,and lead in seawater. Each analysis required 50 µl sample and a 6 minuteanalysis time.

Chromium, Copper, Zinc, Cadmium, Lead, Manganese, Iron, Cobalt, Nickeland Vanadium

Field et al. [747] used ICP high resolution mass spectrometry to determinevanadium, chromium, manganese, iron, cobalt, nickel, copper, zinc, cadmium,and lead in seawater. Only 50 µl samples were required in this analysis.

Nickel, Arsenic, and Vanadium

Alves et al. [744] determined vanadium, nickel, and arsenic in seawater in the10–20 000 ppt range using flow injection cryogenic desolvation ICP-MS.

Beryllium, Aluminium, Zinc, Rubidium, Indium, and Lead

Vandecasteele et al. [745] studied signal suppression in ICP-MS of beryllium,aluminium, zinc, rubidium, indium, and lead in multielement solutions, andin the presence of increasing amounts of sodium chloride (up to 9 g/l). Thesuppression effects were the same for all of the analyte elements under consid-eration, and it was therefore possible to use one particular element, 115indium,as an internal standard to correct for the suppressive matrix effect, whichsignificantly improved experimental precision. To study the causes of matrixeffect, 0.154 M solutions of ammonium chloride, sodium chloride, and caesiumchloride were compared. Ammonium chloride exhibited the least suppressiveeffect, and caesium chloride the most. The results had implications for traceelement determinations in seawater (35 g sodium chloride per litre).

Antimony, Arsenic, and Mercury

Stroh and Voellkopf [746] utilised flow injection analysis coupled to ICP-MSto determine down to 0.6 ppt of antimony, arsenic, and mercury in seawater.

Miscellaneous

Mixtures of metals in seawater have been determined [748, 749] online bysolid-phase chelation and ICP-MS.


Bettinelli and Spezia [750] applied ion chromatography with an ICP-MS tothe determination in seawater of 20 metallic elements in amounts down to1–50 ppt.

The application of ICP-MS to the determination of metals in seawater havebeen reviewed by Bloxam et al. [751].

5.74.10Plasma Emission Spectrometry

Cadmium, Chromium, Copper, Lead, Nickel, and Zinc

Nygaard [752] has evaluated the application of the Spectraspan DC plasmaemission spectrometer as an analysis tool for the determination of trace heavymetals in seawater. Sodium, calcium, and magnesium in seawater are shownto increase both the background and elemental line emission intensities. Op-timum analytical emission lines and detection limits for seven elements arereported in Table 5.8.

Table 5.8. Optimum analysis wavelengths and detection limits for six trace metals in sea-water

Metal Most intense Detection limitEmission line (nm) in seawater (µg/l)

Cd 228.8 5Cr 357.9 1Cu 327.4 2Pb 405.7 16Ni 352.5 6Zn 213.9 3

From [752]


The relative advantages and disadvantages of voltammetric and atomic absorp-tion methodologies are listed below. It is concluded that for laboratories con-cerned with aquatic chemistry of metals (which includes seawater analysis),instrumentation for both AAS (including potentialities for graphite furnaceAAS as well as hydride and cold vapour techniques) and voltammetry shouldbe available. This offers a much better basis for a problem-orientated applica-tion of both methods, and provides the important potentiality to compare thedata obtained by one method with that obtained in an independent manner bythe other, an approach that is now common for the establishment of accuracyin high-quality trace analysis.


Voltammetry

Advantages DisadvantagesSimultaneous analysis for severalelements per run

Prior photolytic decomposition ofdissolved organic matter required frommany types of sample including seawater

Substance specific Suspended particulates need priordigestion

Suitable for speciation studies Applicable to limited rangeofmetals, e.g.,Cu, Pb, Cd, Zn, Ni, Co

3–4 elements per hour

Graphite Furnace Atomic Absorption Spectrometry

Advantages DisadvantagesHigh analysis rate Nonspecific absorption3–4 elements per hour Spectral interferencesApplicable to many more metalsthan voltammetric methods

Element losses by molecular distillationbefore atomisation

Superior to voltammetry formercury and arsenic particularlyin ultratrace range

Limited dynamic range

Contamination sensitivityElement specific (or one element per run)Not suitable for speciation studies inseawaterPrior separation of sea salts from metalsrequiredSuspended particulates need priordigestionAbout three times as expensive asvoltammetric equipmentInferior to voltammetry for cobalt andnickel

Anodic Stripping Voltammetry

Earlier work on the application of anodic stripping voltammetry to the deter-mination of methods in seawater is reviewed in Table 5.9.


Table 5.9. Metals in seawater – anodic stripping voltammetry

Zn, Cd, Pb, Cu 1–10 nm mol/l [753]Zn, Cd, Pb, Cu Zn [754]

Pb 0.01–0.1 mg/lCuCd 0.001–0.1 mg/l [755]

Cd, Pb Cd 0.18 µg/lPb 0.21 µg/l

Bi, Cu, Pb, Cd, Zn – [756]Cu, Pb, Cd, Zn – [757]Tl 0.2–1 µmol/l [758]Pb, Cd 1 µg/l [759]Pb 4 µg/l [760]

Source: Author’s own files

A variety of electrodes have been used in this technique, including rotat-ing mercury coated vitreous carbon [756], wax impregnated graphite cylin-ders [758], and hanging mercury drops [753–755].

It was the development of the rotating glassy carbon electrode with a pre-plated or co-plated mercury film that gave this technique the sensitivity andresolution required for use in seawater.

Anodic stripping voltammetry using a rotating glassy carbon electrode hasbeen extensively used to study metal organic interactions. The instrumenta-tion is adaptable to use at sea and does not generally require any chemicalpretreatment of samples prior to analysis. This permits rapid analysis at in situpH, thus minimising any changes in speciation due to storage [310, 756–764].

A typical electrode tip consists of a 6 mm glassy carbon disk sealed withTeflon, that had been polished with diamond polishing compound. The ref-erence electrode is an Ag/AgCl type inserted into an acid-cleaned Vycor tipTeflonbridge tube containing clean seawater.Highpurity argonwhich ispassedthrough a high-temperature catalytic scrubber to remove oxygen and then isrehumidified using an in-line natural seawater bubbler prior to use is com-monly used as a purge gas. Electronic interfaces between the polarograph andthe electrode permit the polarograph to control all steps of the analysis auto-matically, including purging. This sample is placed in an acid-cleaned Teflonpolarography cell that has been copiously rinsed with the sample to be anal-ysed. Samples are analysed as quickly as possible after collection at naturalpH for free and other easily reducible species, using a pre-plated mercury-filmtechnique.

Because of differing sensitivities and the natural levels of free metal or theanodic scanning voltammetric labile metal, cadmium, and copper in seawa-ter are analysed using a 10 minute plating time, a –1.0 V plating potential,and scanning in 6.67 mV/s increments. Zinc determinations can be made ona fresh aliquot of sample to eliminate any possible effects due to Cu–Zn inter-


metallic complex formation. Zinc is analysed by plating at –1.25 V for 5 min.The remaining operating conditions are the same as described previously forcopper and lead. The detection limits of this system were approximately: zinc0.02 nmol/kg, cadmium 0.02 nmol/kg, and copper 0.3 nmol/kg.

Heavy Metals

Although anodic stripping voltammetry is one of the few techniques suitablefor the direct determination of heavy metals in natural waters [310,756–764], itis not readily adaptable to in situ measurements. Lieberman and Zirino [623]examined a continuous flow system for the anodic stripping voltammetry de-terminationof zinc in seawater, usinga tubular graphite electrodepredepositedwith mercury. A limitation of the approach was the need to pump seawater tothe measurement cell, while the method required the removal of oxygen withnitrogen before measurements.

Batley and Matousek [390, 778] examined the electrodeposition of the ir-reversibly reduced metals cobalt, nickel, and chromium on graphite tubes formeasurement by electrothermal atomisation. This method offered consider-able potential for contamination-free preconcentration of heavy metals fromseawater. Although only labile metal species will electrodeposit, it is likely thatthis fraction of the total metal could yet prove to be the most biologicallyimportant at the natural pH [779].

Batley [780] examined the techniques available for the in situ electrode-position of lead and cadmium in seawater. These included anodic scanningvoltammetry at a glass carbon thin film electrode and the hanging drop mer-cury electrode in the presence of oxygen, and in situ electrodeposition onmercury-coated graphite tubes.

Batley [780] foundthat in situdepositionof leadandcadmiumonamercury-coated tube was the more versatile technique. The mercury film, deposited inthe laboratory, is stable on the dried tubes which are used later for field elec-trodeposition. The deposited metals were then determined by electrothermalAAS.

Scarponi et al. [781] studied the influence of an unwashed membrane fil-ter (Millpore type HA, 47 mm diameter) on the cadmium, lead, and copperconcentrations of filtered seawater. Direct simultaneous determination of themetals was achieved at natural pH by linear-sweep anodic stripping voltam-metry at a mercury film electrode. These workers recommended that at least1 litre of seawater be passed through uncleaned filters before aliquots for anal-ysis are taken; the same filter can be reused several times, and only the first50–100 ml of filtrate need be discarded. Samples could be stored in polyethy-lene containers at 4 ◦C for three months without contamination, but losses oflead and copper occurred after five months of storage.

Brugmann et al. [782] compared results obtained by ASV and AAS in thedetermination of cadmium, copper, lead, nickel, and zinc in seawater. Three


methods were compared. Two consisted of AAS but with preconcentrationusing either freon or methyl isobutyl ketone, and ASV was used for cadmium,copper, and lead only. Inexplicable discrepancies were found in almost allcases. The exceptions were the cadmium results by the two methods and thelead results from the freon with AAS methods and the ASV methods.

Clem and Hodgson [783] discuss the temporal release of traces of cadmiumand lead in bay water from EDTA, ammonium pyrrolidine diethyldithiocarba-mate, humic acid, and tannic acid after treatment of the sample with ozone.Anodic scanning voltammetry was used to determine these elements.

Nygaard et al. [752] compared two methods for the determination of cad-mium, lead, and copper in seawater. One method employs anodic strippingvoltammetry at controlled pH (8.1, 5.3 and 2.0); the other involves sample pre-treatment with Chelex 100 resin before ASV analysis. Differences in the resultsare discussed in terms of the definition of available metal and differences inthe analytical methods.

Decreasing the pH makes more metal available to the electrode during theplating period. The metals are apparently made available through a number ofprocesses:

(1) Protonation of inorganic complexing anions such as carbonate, bicarbon-ate, sulfate, and hydroxide

(2) Dissolution of gelatinous hydrous iron oxide, which adsorbs and occludesmetals

(3) Protonation of organic complexing agents

Brugmann [784] discussed different approaches to trace metal speciation(bioassays, computer modelling, analytical methods). The electrochemicaltechniques include conventional polarography, ASV, and potentiometry. ASVdiagnosis of seawater was useful for investigating the properties of metal com-plexes in seawater. Differences in the lead and copper values yielded for Balticseawater by methods based on differential pulse ASV or AAS are discussedwith respect to speciation.

Bruland et al. [785] compared voltammetric and AAS (with preconcentra-tion) methods in the determination of copper, lead, and cadmium in seawater.

Cyclohexane-1,2-dione dioxime (nioxime) complexes of cobalt (II) andnickel (II) were concentrated from 10 ml seawater samples onto a hangingmercury drop electrode by controlled adsorption. Cobalt (II) and nickel (II)reduction currents were measured by differential pulse cathodic strippingvoltammetry. Detection limits for cobalt and nickel were 6 pM and 0.45 mM,respectively.The results ofdetailed studies foroptimising theanalytical param-eters, namely nioxime and buffer concentrations, pH, and adsorption potentialare discussed.

Achterberg, Van den Berg, and others [786] used a voltammetric techniqueto take continuous real-time measurements of nickel, copper, and zinc in theIrish Sea.


Cuculic and Branica [788] used differential pulse ASV to study the ad-sorption of cadmium, lead, and copper on glass, quartz, and Nalgene samplecontainers. Nalgene was shown to be the best for sample storage, and quartzthe best for electroanalytical vessels.

Bond et al. [791] studied strategies for trace metal determination in seawa-ter by ASV using a computerised multi-time domain measurement method.A microcomputer-based system allowed the reliability of the determinationof trace amounts of metals to be estimated. Peak height, width, and poten-tial were measured as a function of time and concentration to construct thedatabase. Measurements were made with a potentiostat polarographic analyserconnected to the microcomputer and a hanging drop mercury electrode. Thepresence of surfactants, which presented a matrix problem, was detected viatime domain dependent results and nonlinearity of the calibration. A decisionto pretreat the samples could then be made. In the presence of surfactants, nei-ther a direct calibration mode nor a linear standard addition method yieldedprecisedata.Alternativeways to eliminate the interferencesbasedeitheron the-oretical considerations or destruction of the matrix needed to be considered.

Cadmium, Copper, Lead, Antimony, and Bismuth

Brihaye et al. [787] have described a procedure for the determination of theseelements in seawater.

Results obtained for cadmium, lead, copper, antimony, and bismuth inseawater by two different methods (linear ASV with a ring-disk electrode anddifferential pulse anodic scanning voltammetry with a hanging mercury dropelectrode) were in good agreement for UV-irradiated samples. Linear anodicscanning voltammetry with the ring-disk electrode gave systematically highercadmium, lead, and copper contents because of exchange of mercury (II) ionsadded to the solution, with the heavy metal non-labile complexes.

Copper and Mercury

Sipos et al. [789] have described a procedure for the simultaneous determina-tionof copper andmercury in seawaterdown to theng/l rangeusingdifferentialpulse ASV at a gold electrode. Pretreatment is necessary, and comprises UVirradiation to release the trace metal bound to dissolved organic matter.

A relative standard deviation of 2.7% was obtained for copper at the 0.35 µg/kg level, and 18.6% at the 0.026 µg/kg for mercury.

Miscellaneous

Bott [790] reviewed of voltammetric methods for the determination of tracemetals in seawater and other natural waters.

Nurnberg [792] has studied in great detail various aspects which are im-portant to obtaining reliable results by voltammetric methods. These include


sampling, sample pretreatment steps, optimum pH adjustment, sample stor-age, decomposition of dissolved organic matter, the voltammetric determina-tion, digestion of filtered-off suspended matter and automation. These detailsare extremely interesting for anyone who is considering setting up voltammet-ric methods in their laboratory. Nurnberg [792] also discusses results obtainedby these techniques on seawater samples from the Mediterranean and Belgian,Dutch, and German coastal zones of the North Sea, Norwegian Sea and NorthAtlantic, the Pacific, the Arctic Oceans, and the Weddell Sea.

Differential Pulse Anodic Stripping Voltammetry

Duinker and Kramer [311] studied the speciation of dissolved zinc, cadmium,lead, and copper in North Sea water by differential pulse ASV.

Dissolved electroactive concentrations of zinc, cadmium, lead, and copperin the North Sea samples were measured at natural and lower pH values by thistechnique using a Kemula-type hanging mercury drop.

Studies with spiked seawater showed that low concentrations of cadmiumand lead could be measured in the presence of oxygen by using differentialpulse ASV at the hanging mercury drop electrode [780]. The presence ofoxygen resulted in a highly sloping baseline giving rise to greater analyticalerrors. In samples buffered to pH 4.8, peak heights and peak potentials didnot differ significantly before and after oxygen removal. For samples at thenatural pH of 7.8, although the cadmium wave was unchanged, the lead wavein the presence of oxygen was greater in height by 21% and shifted by 15 mVto a more negative potential.

At the glassy carbon electrode, using both in situ and preformed mercuryfilms, similar results were obtained, but the sloping baseline interference ob-served at the hanging mercury drop electrode was less evident because of thehigher stripping currents.

The in situ electrodeposition technique was applied to the determinationof lead in saline waters of the Port Hacking Estuary near Sydney. Graphitetubes precoated with mercury were used in the immersible Perspex electrodeprobe. For natural lead concentrations, depositions in excess of 15 minuteswere required to give absorbance values greater than 0.1 during atomisation.Blank values for the coated-tubes were low, but increased for tubes immersedin the sampled water at a controlled potential below that required for leaddeposition (–0.3 V versus Ag/AgCl), for deposition times similar to those usedfor lead determinations. Results for lead showed good agreement with thosefor labile lead determined independently by stripping voltammetry.

The limits of detection for metals in seawater using in situ graphite tubeelectrodeposition will be governed by the deposition time. Unlike laboratoryanalyses, there will be no depletion of metals from the solution when lengthydeposition times are used, since fresh sample is being continuously pumpedthrough the electrode. It should therefore be possible to detect the extremely


low metal concentrations in open ocean water. For lead it was found, forexample, that a 2 h deposition in the presence of oxygen gave a measuredlead atomisation absorbance equivalent to twice the blank value, for seawatercontaining 10 ng Pb per litre.

Average concentrations detected in North Sea samples at salinities ≥ 32% Sand their ranges are (in µg/l): 3.9 (2.0–7.5) for zinc, 0.23 (0.13–0.31) forcadmium, 0.3 (0.1–0.6) for lead, and 0.3 (0.25–0.60) for copper (pH 8.1).The ammonium pyrrolidine dithiocarbamate methylisobutyl ketone extrac-tion/concentration method, followed by AAS measurement applied to the samesamples, resulted in concentrations (in µg/l) of 3.9 (2.0–7.5) for zinc, 0.11(0.01–0.27) for cadmium, 0.5 (0.2–0.9) for lead, and 1.6 (0.7–3.2) for copper.A fraction of the electroactive concentrations at pH 2.7 (6.1 for zinc) areelectroactive at pH 8.1. The fractions are 100% for cadmium, 20% for copper,13% for lead, and 40% for zinc. The remaining fractions are partly composedof organically bound species in solution. The low value for lead may resultfrom the presence of particulate lead dissolved at low pH.

North Sea water at the natural pH has a complexing capacity, probably dueto the presence of dissolved organic compounds, in a concentration equiva-lent to 0.3 M copper. The complexing capacity is zero at pH 2.7. The methodof standard addition for the determination of electroactive copper and leadconcentrations may lead to erroneous results in samples where complexationof this type occurs.

Cuculic and Branica [788] applied differential pulse anodic stripping vol-tammetry to a study of the adsorption of cadmium, copper, and lead in seawateronto electrochemical glass vessels, quartz cells, and Nalgene sample bottles.Nalgene was best for sample storage and quartz was best for electroanalyticalvessels.

For most tasks in the trace chemistry of natural waters, voltammetric de-termination requires preconcentration, because in a group of simultaneouslydetermined ecotoxic heavy metals one usually has levels below 0.1 µg/l.

Electrochemical preconcentration can be achieved in the following two dif-ferent ways, depending on whether differential pulse stripping voltammetry(differential pulse ASV) or adsorption differential pulse voltammetry has beenapplied.

Heavy metals capable of forming amalgams, e.g., Cu, Pb, Cd, Zn, etc., areplated at a stationary mercury electrode consisting of a hanging mercury dropelectrode with adjustment of a rather negative potential of –1.2 V versus theAg/AgCl reference electrode for several minutes. To speed up mass transfer,the solution is stirred with a magnetic bar at 900 rpm. Their preconcentrationis achieved by the accumulation of the heavy plated metals in the mercurydrop. Subsequently the stirring is terminated, and after a quiescent period of30 seconds the potential is scanned in the anodic direction in differential pulsemode. At the respective redox potential the plated heavy metal is reoxidisedand the corresponding current is recorded (Fig. 5.22). The voltammetric peak


heights obtained are proportional to the bulk concentrations of the respectivetrace metals in the analyte.

The hanging drop mercury electrode can usually be applied down to tracelevels of 0.1–0.05 µg/l.

At lower ultratrace levels, the less voluminous mercury film electrode hasto be used. It consists of a mercury film of only several hundred nm thicknesson a glassy carbon electrode as support. The fabrication of this glassy carbonelectrode is critical for obtaining an optimal mercury film electrode suitableto perform determinations down 1 ng/l or below.

Turner et al. [331] showed that rapid staircase stripping at 1–2 ms stepwidths provides a fast sensitive alternative to differential pulse stripping forfield use, particularly in the automated determination of copper.

Mart et al. [793] and Valenta et al. [794] have described two differentialpulse ASV methods for the determination of cadmium, lead, and copper inarctic seawater. After a previous plating of the trace metals into a mercury filmon a rotating electrode with a highly polished glassy carbon as substrate, theywere stripped in the differential pulse mode. The plating was done in situ.

The film formed during the blank test was left on the glassy carbon surface.Upon addition of further mercury nitrate, 20–50 µ1 of a 5 g/l solution were

Figure 5.22. Voltammogram of the simultaneous determination of Cu, Pb, Cd, and Zn withDPASV at the HMDE, and subsequent determination of Se1v by DPCSV in the same run inrain water at an adjusted pH of 2. Preconcentration time for DPASV 3 min at –1.2 V, forDPCSV 5 min at –0.2 V. 1 Original analyte. 2 After first standard addition. Total analysistime with two standards. Source: Author’s own files


added to 50 ml. Analysis was then performed. For samples where copper wasexpected to be close to the determination limit of 10 ng/kg, only the film platedduring the blank test run was used, thus avoiding an increasing slope due togrowth of the mercury film.

Pihlar et al. [795] have described a sensitive differential pulse voltammetricmethod for the determination of nickel and cobalt in seawater. The pH of thesample is adjusted to 9.2–9.3 by adding an ammonia–ammonium chloridebuffer. Optimal ammonia buffer concentration is 0.1 M for nickel concentra-tionsbelow10 µg/kg, and20 µgof a0.1 Mdimethylglyoximesolution inethanolis added to a 50 ml sample. The analyte is de-aerated for 10 minutes. At theworking electrode, a hanging mercury drop electrode, a potential of –0.7 V isadjusted and the nickel–dimethylglyoxime complex is adsorbed at the mercurysurface. To speed up mass transfer the solution is stirred with a magnetic bar.Depending on the concentration of nickel (and cobalt), 5–10 min of adsorptiontime are needed. After a rest period of 30 s the voltammogram is recorded byscanning the potential in the negative direction. Concentrations were evaluatedby standard addition.

Other groups of elements that have been determined in seawater by differen-tial pulse stripping voltammetry include cadmium, copper, and zinc [796]; cop-per, lead, and cadmium [797]; and zinc, cadmium, lead, and copper [773,798].

Krznaric [799] studied the influence of surfactants (EDTA, NTA) on mea-surements of copper and cadmium in seawater by differential pulse ASV. Ad-sorption of surfactants onto the electrode surface were shown to change thekinetics of the overall electrode charge and mass transfer, resulting in altereddetection limits. Possible implications for studies on metal speciation in pol-luted seawater with high surfactant contents are outlined.

Andruzzi et al. [800, 801] discussed the use of a long-lasting sessile-dropmercury electrode in the differential pulse ASV subtrace determination of zinc,cadmium, and lead in seawater. A repeatability of about 3% and a detectionlimit of 10–10 M were achieved for these three metals.

Adsorption Differential Pulse Voltammetry

A number of heavy metals are not capable of forming stable amalgams [859].Adsorption of suitable species has to be utilised for their electrochemicalpreconcentration at the electrode interface. Thus, for example, in the determi-nation of copper and nickel, the sample is adjusted to pH 9.2 with ammoniabuffer and dimethylglyoxime added to form the copper and nickel chelates.Dimethylglyoxime transforms a certain amount into the chelates Ni(DMG)2,while another fraction of the overall concentrations of Ni and Co exists asammonia complexes in the analyte. Then a potential of –0.7 V is adjusted forseveral minutes at the hanging mercury drop electrode. This potential in therange of the zero-charge potential of the mercury electrode is most favourablefor the adsorption of the dimethylglyoxime chelates of both heavy metals. To


speed up mass transfer, the solution is stirred at 900 rpm during the adsorp-tion time. The adsorbed amount of the chelates is proportional to their bulkconcentration. As all the complex equilibria in the analyte between the vari-ous complexes of nickel and cobalt with ammonia and dimethylglyoxime asligands are adjusted, the adsorbed amount of their dimethylglyoxime chelatesis also proportional to the total bulk concentrations of both heavy metals inthe analyte. The proviso is that the adsorbed amounts of the dimethylglyoximechelates should correspond to the rising part of the adsorption isotherms, andfull coverage of the electrode surface is avoided. With this approach, substan-tial preconcentration is attained at the electrode surface within adsorptiontimes of a few minutes. Then the electrode potential is made more negative inthe differential pulse mode until the reduction potentials of nickel and cobaltare reached, and the peaks produced by the reduction of the adsorbed chelatespecies Ni(DMG)2 and Co(DMG)2 are recorded.

The ultimate determination limits obtainable by voltammetric methods arein the range 0.02–0.1 ppb. It is emphasised that these are practical limits.Minimisation of the blanks would enable still lower determination limits to beachieved.

Potentiometric Stripping Analyser

This method in some ways resembles the technique for ASV [321,322]. The ana-lyticaldevice isbasedona three-electrode system: (1) aglassycarbonelectrode,which serves as a cathode; (2) a saturated calomel electrode (SCE), which is thereference electrode; and (3) a platinum counter-electrode during electrolysis.

Analysis of metal ions in a sample solution is started by electrochemicalformation of a mercury film on the glassy carbon electrode. Subsequently,the metal ions are reduced and amalgamated in the mercury film during theelectrolysis step (plating). When the plating is terminated, the metals arestripped from the mercury film back into the solution by chemical oxidation.During this step the potential of the carbon electrode (relative to the SCE) isrecorded as a function of time. The metals are identified by their strippingpotentials, and the quantitative determination is obtained by measuring thestripping time for each metal.

Jagner et al. [802] used this technique to determine zinc, cadmium, lead, andcopper in seawater. Their method includes computer control of the potentio-metric stripping technique. They compared their results with those obtainedby solvent extraction–AAS and showed that the computer-controlled poten-tiometric stripping technique is more sensitive, and has advantages over ASV.Computer control makes deoxygenation of the sample unnecessary.

Water samples from the Arctic Sea were analysed by the potentiometricstripping technique. Lead (II) and cadmium (II) were determined after pre-electrolysis for 32 min at –1.1 V with respect to Ag/AgCl; the detection limitswere 0.06 and 0.04 nmol/l, respectively. Zinc (II) was determined after the ad-


Figure 5.23. Potentiometric stripping curve for a 25 ml seawater sample. I Background.II Curve before standard addition. III and IV Curve after each addition of 5 ng CdII and100 µg PbII. Internal standard: 25 ng CuII. Plating time 32 min at 95 V relative to SCE.Source: [803]

dition of gallium (III) by pre-electrolysis for 16 min at –1.4 V with respect toAg/AgCl; the detection limit was 0.25 nmol/l. Problems in the determination ofcopper (II) at the very low concentrations found in oceanic waters are outlined.The average zinc (II), cadmium (II), and lead (II) concentrations in eight dif-ferent samples were 2.5, 0.16, and 0.10 nmol/l as determined by potentiometricstripping analysis, and 1.9, 0.16, and 0.09 nmol/l as determined by solventextraction–AAS. The advantages of this computer-controlled technique for theanalysis of seawater are discussed.

Drabek et al. [803] applied potentiometric stripping analysis to the deter-mination of lead, cadmium, and zinc in seawater. The precision was evaluatedby several duplicate determinations and was found to be in the range 5–16%relative, depending on the concentration level. The accuracy of the methodwas evaluated by comparison with other conventional methods, e.g., AAS andASV, and good agreement between the methods was found.

Figure 5.23 shows typical stripping curves for cadmium, lead, and zincobtained from a 25 ml seawater sample. The sample was analysed as previouslydescribed. The concentrations found were 7.1 µg/l (PbII), 0.2 µg/l (CII), and4.1 µg/l (ZnII).


Heavy Metals

Donat and Bruland [804] studied the speciation of copper and nickel in sea-water by competitive ligand equilibration–cathodic stripping voltammetry,differential pulse ASV, and graphite furnace AAS.

Donat and Bruland [217] conducted a direct determination of cobalt andnickel in seawater by differential pulse cathodic scanning voltammetry, pre-ceded by absorptive collection of cyclohexane-1,2 dioxime complexes.


Perez-Pina et al. [805] studied the use of triethanolamine and dimethyl-glyoxime complexing agents in absorptive cathodic stripping voltammetry ofcobalt and nickel in seawater. Nickel and cobalt could be determined at levelsdown to 2 nM and 50 pM, respectively.

Huynk et al. [218] also used differential pulse cathodic stripping voltam-metry for the determination of cobalt and nickel in seawater by dimethylgly-oxime complexation. They report detection limits of 0.002 µg/l for cobalt and0.005 µg/l for nickel.

Abollino et al. [690] compared cathodic stripping voltammetry and graphitefurnace AAS in determination of cadmium, copper, iron, manganese, nickel,and zinc in seawater. The effects of UV irradiation, acidification, and onlinesample preconcentration were studied.

Heavy Metals, Uranium, Vanadium, Molybdenum

Cathodic stripping voltammetry has been used [807] to determine lead, cad-mium, copper, zinc, uranium, vanadium, molybdenum, nickel, and cobalt inwater, with great sensitivity and specificity, allowing study of metal specia-tion directly in the unaltered sample. The technique used preconcentration ofthe metal at a higher oxidation state by adsorption of certain surface-activecomplexes, after which its concentration was determined by reduction. Thereaction mechanisms, effect of variation of the adsorption potential, maxi-mal adsorption capacity of the hanging mercury drop electrode, and possibleinterferences are discussed.

Selenium

Certain trace substances such as SeIV can be determined in similar fashionby differential cathodic stripping voltammetry (DPCSV) [895]. For seleniuma rather positive preconcentration potential of –0.2 V is adjusted. SeIV is re-duced to Se2– and Hg from the electrode is oxidised to Hg2+ at this potential.It forms, with Se2– on the electrode, a layer of insoluble HgSe and in this man-ner the preconcentration is achieved. Subsequently the potential is altered inthe cathodic direction in the differential pulse mode. The resulting HgII peakproduced by HgII reduction is proportional to the bulk concentration of SeIV

in the analyte.

5.74.13Chronopotentiometry

Nickel and Cobalt

Eskilsson et al. [868] have described equipment for automated determinationof traces of cobalt and nickel by potentiometric stripping analysis, which useda freshly prepared mercury film on a glassy carbon support as the working


electrode. The use of concentrated electrolyte (5 M calcium chloride) providedvirtually oxygen-free stripping solutions. Analytical procedures for aqueoussolutions and biological samples are described.


Heavy Metals

Knochel andPrange [812] convertedmetals in seawater into theirdiethyldithio-carbamates prior to X-ray fluorescence analysis of the separated solids. Mem-brane filtration of the precipitates resulted in carbarmate-loaded filters, whichcould be directly measured using radioisotope-excited X-ray fluorescence anal-ysis. Furthermore, elution of Chromosorb columns loaded with the dithiocar-bamate complex by the passage of chloroform yielded chloroform solutionsin which the trace metals could be determined by X-ray fluorescence analysisusing totally reflecting sample supports. Similarly, the precipitate on the mem-brane filter could be dissolved in chloroform and determined in the resultingsolution. The sensitivity of this method and the pH dependence of the reactionwere also investigated.

Prange et al. [813] give details of equipment and a procedure for determi-nation of traces of heavy metals by solvent extraction using di-isobutyl ketoneand isobutyl methyl ketone, combined with microdroplet analysis by X-rayfluorescence spectrometry using a specially designed filter paper; sodium di-ethyldithiocarbamate is used as chelating agent. The limits of detection formanganese, iron, cobalt, nickel, copper, zinc, and lead were 15, 16, 8, 8, 13, 13,and 40 µg/l, respectively, for a 100 µl sample volume. The method was appliedto analyses of seawater from Chirihama, Japan. Results obtained are in fairagreement with the reference values obtained by AAS.

Heavy Metals, Vanadium

Morris [814] separated microgram amounts of vanadium, chromium, man-ganese, iron, cobalt, nickel, copper, and zinc from 800 ml of seawater by pre-cipitation with ammonium tetramethylenedithiocarbamate, and extraction ofthe chelates at pH 2.5 with methylisobutyl ketone. Solvent was removed fromthe extract, the residue was dissolved in 25% nitric acid, and the inorganicresidue was dispersed in powdered cellulose. The mixture was pressed intoa pellet for X-ray fluorescence measurements. The detection limit was 0.14 µgor better when a 10 min counting period was used.

Heavy Metals, Uranium

Preconcentration by the ring-oven technique [227] and by solvent extractionhave both been used in order to improve the sensitivity of X-ray fluorescence


spectrometry. Armitage and Zeitlin [227] converted uranium, copper, nickel,cobalt, iron, and manganese to the 8-hydroxyquinolates and extracted thesewith chloroform. The extract was applied to a filter paper disk in a ring ovenat 160 ◦C, and the metals were separated prior to final determination by theX-ray technique.

Heavy Metals, Vanadium, Molybdenum, Mercury, Uranium, Potassium, Calcium,Titanium, Gallium, Arsenic, Selenium, Rubidium, Strontium, Yttrium, Zirconium,Silver, Antimony, and Barium

Prange et al. [809, 810] carried out multielement determinations of the stateddissolved heavy metals in Baltic seawater by total reflection X-ray fluorescence(TXRF) spectrometry. The metals were separated by chelation adsorption ofthe metal complexes on lipophilised silica-gel carrier and subsequent elution ofthe chelates by a chloroform/methanol mixture. Trace element loss or contami-nation could be controlled because of the relatively simple sample preparation.Aliquots of the eluate were then dispersed in highly polished quartz samplecarriers and evaporated to thin films for spectrometric measurements. Recov-eries (see Table 5.10), detection limits, and reproducibilities of the method forseveral metals were satisfactory.

This analytical method, based on TXRF, enables a large number of traceelements to be determined simultaneously. The range is suitable for differ-ent areas of the sea. The motivation to use TXRF resulted mainly from thecharacteristic features of the method: its high detection power, its universalcalibration curve, which eliminates the need for matrix-dependent standardsamples or standard-addition procedures, the simple preparation of the samplefilms, and of course the possibility of multielement determination.

Haarich et al. [811] applied total reflection X-ray fluorescence spectrometryto the determination of a wide range of miscellanous metals in seawater.


The application of neutron activation analysis to water samples has severaladvantages. Like inductively coupled plasma atomic absorption spectrometryit is a multielement analysis technique, and as such it is useful for performingelement scans when looking for unknown elements. It enables several elementsto be determined in the same run. It is not limited to metallic elements. Itsdisadvantage is high cost and the probability that most laboratories cannothouse the reactor equipment, so that in fact samples must be sent away foranalysis to an establishment with such facilities. This delays results and de-mands careful consideration of sample preservation during the waiting period.The technique is intrinsically sensitive, and can be made more so by precon-centrating the metals onto a resin such as Chelex 100, which can be directly


Tabl

e5.

10.

Rec

over

ies

dete

rmin

edby

stan

dard

addi

tion

tose

awat

erw

ith

APD

Cor

NaD

BD

TC

asth

ech

elat

ing

agen

t

Ele

men

tE

lem

entf

ound

(ng/

kg)

wit

hA

PDC

Ele

men

tfou

nd(n

g/kg

)w

ith

NaD

BD

TC

Add

itio

nals

tand

ard

(ng/

kg)

Bef

ore

addi

tion

Aft

erad

diti

onY

ield

(%)

Bef

ore

addi

tion

Aft

erad

diti

onY

ield

(%)

V25

0014

8±

6118

8±

79–

1540

±10

039

20±

7095

.3±

5.2

Mn

1000

––

–32

±6

917

±9

88.5

±1.

2Fe

1000

353

±51

1320

±43

96.3

±4.

834

4±

1713

70±

2010

2.8

±2.

9C

o10

00in

tern

alst

anda

rd10

0in

tern

alst

anda

rd10

0N

i10

0016

5±

1573

5±

2157

.0±

6.9

241

±12

1230

±20

99.2

±2.

4C

u10

0093

±23

790

±17

69.7

±6.

110

7±

210

90±

2098

.2±

2.3

Zn

1000

64±

1214

6±

188.

2±

6.4

179

±3

1190

±23

100.

8±

2.3

Se25

00–

1970

±10

078

.8±

5.1

–25

30±

3010

1.0

±3.

1M

o50

0055

90±

240

9210

±17

072

.3±

9.8

1020

0±

490

1520

0±

670

100.

2±

5.5

Cd

1000

–31

3±

2331

.3±

2.3

–98

4±

2598

.4±

2.5

Hg

1000

196

±27

925

±44

72.9

±6.

175

±10

1050

±30

97.6

±2.

9Pb

1000

35±

747

2±

3143

.7±

5.2

49±

411

00±

4010

5.2

±3.

8U

2500

557

±51

772

±65

8.6

±3.

131

00±

190

5650

±11

010

2.1

±8.

6

N=

3Fr

om:[

809]

APD

C:a

mm

oniu

mpy

rrol

iden

edi

thio

carb

amat

eN

aDB

DT

C:s

odiu

mdi

benz

yldi

thio

carb

amat

e


Table 5.11. Application of neutron activation analysis to the determination of metals inseawater

Elements Sample pre-concentration Reference

Co, Cr, Ca, Fe, Pb, Se, Sr Sea salts Freeze dried [817]As, Cu, Sb Thionalide precipitation [818]Misc. elements Frozen sample [819]Al, V, Cu, Mo, Zn, U Organic co-precipitants

(8-hydroxyquinoline)[21]

Ag, Au, Cd, Ce, Co, Cr, Eu, Fe, Hg,La, Mo, Sc, Se, U, Zn, As, Sb

Adsorption and charcoal [628]

Ba, Ca, Cd, Ce, Co, Cr, Cu, Fe, La, Mg,Mn, Sc, U, V, Zn

Adsorption of Chelex-100 glass powder [627]

Hg, Au, Cu Co-precipitation with leaddiethyldithiocarbamate

[820]

12 elements Chelating resin [821]Transition metals Chelating resin [822]As, Mo, U, V Colloid flotation [823]Th Ion exchange chromatography [889]Co, Cu, Hg Co-precipitation with lead pyrrolidine

dithiocarbamate[824]


analysed by the neutron activation technique. With these facilities, analysis formany metals at the ultra-low background levels at which they occur in seawaterbecomes a possibility.

Some applications of neutron activation analysis to seawater analysis aresummarised in Table 5.11.

The application of neutron activation techniques to the measurement oftrace metals in the marine environment has been reviewed by Robertson andCarpenter [815, 816].

Heavy Metals

Stiller et al. [824] have described the determination of cobalt, copper, andmercury in Dead Sea water by neutron activation analysis followed by X-rayspectrometry and magnetic deflection of β-ray interference.

The metals were coprecipitated with lead–ammonium pyrrolidine dithio-carbamate and detected by X-ray spectrometry following neutron activation.Magnetic fields deflect the β rays while the X rays reach the silicon (lithium)detector undeviated. The detectors have low sensitivity to γ rays. The con-centration of cobalt found by this method was 1.3 µg/l, about one-fifth of thatmeasuredpreviously,while thatof copper, 2.0 µg/l, agreedwith results obtainedby some previous workers. The concentration of mercury was 1.2 µg/l.


Arsenic, Molybdenum, Uranium, and Vanadium

Murthy and Ryan [823] used colloid flotation as a means of preconcentrationprior to neutron activation analysis for arsenic, molybdenum, uranium, andvanadium. Hydrous iron (III) oxide is floated in the presence of sodium decylsulfate with small nitrogenbubbles from 1 litre of seawater at pH5.7. Recoveriesof arsenic, molybdenum, and vanadium were better than 95%, whilst that ofuranium was about 75%.

Silver, Chromium, Cadmium, Copper, Manganese, Thorium, Uranium, and Zirconium

Holzbecker and Ryan [825] determined these elements in seawater by neutronactivation analysis after coprecipitation with lead phosphate. Lead phosphategives no intense activities on irradiation, so it is a suitable matrix for trace metaldeterminations by neutron activation analysis. Precipitation of lead phosphatealso brings down quantitatively the insoluble phosphates of silver (I), cad-mium (II), chromium (III), copper (II), manganese (II), thorium (IV), ura-nium (VI), and zirconium (IV). Detection limits for each of these are given,and thorium and uranium determinations are described in detail. Gammaactivity from 204Pb makes a useful internal standard to correct for geometrydifferences between samples, which for the lowest detection limits are countedclose to the detector.

Cadmium, Cobalt, Chromium, Copper, Iron, Manganese, Molybdenum, Nickel,Scandium, Tin, Thorium, Uranium, and Zinc

Greenberg and Kingston [821,822] used a solid Chelex 100 resin to preconcen-trate these elements from 100–500 ml of estuarine and seawater prior to theirdetermination. A procedure is described for the preconcentration of 100 ml ofestuarine and seawater into a solid sample using Chelex 100 resin. This solidsample weighs less than 0.5 g, and contains the transition metals and manyother elements of interest, but is essentially free of alkali metals, alkaline earthmetals, and halogens.

The application of the Chelex 100 resin separation and preconcentration,with the direct use of the resin itself as the final sample for analysis, is anextremely useful technique. The elements demonstrated to be analyticallydeterminable from high salinity waters are cobalt, chromium, copper, iron,manganese, molybdenum, nickel, scandium, thorium, uranium, vanadium,and zinc. The determination of chromium and vanadium by this techniqueoffers significant advantages over methods requiring aqueous final forms, inview of their poor elution reproducibility. The removal of sodium, chloride,and bromide allows the determination of elements with short and intermedi-ate half-lives without radiochemistry, and greatly reduces the radiation dosereceived by personnel. This procedure was successfully applied in a study of


more than 100 samples collected throughout the entire length of the Chesa-peake Bay. The salinity of these samples varied from that of fresh water to thatof Atlantic Ocean water.

Thorium

Huh and Bacon [589] used neutron activation analysis to determine 232Th inseawater. Seawater samples were subjected to pre- and post-irradiation pro-cedures. Separation and purification of the istopes, using ion exchange chro-matography and solvent extraction, were performed during pre-irradiation.After irradiation, 233Pa was extracted and counted. Yields were monitored with230Th and 231Pa tracers. Measured 232Th concentrations were 27 ×10–7 dpm/kgfor deep-water samples from below 400 m.

Multiple Elements

In a procedure [627] employing preconcentration of the metals on a columncontaining a mixture of Chelex 100 and Pyrex glass powder, the problems as-sociated with swelling of the Chelex 100 were overcome and constant flow ratesof sample down the column achieved. The water samples were passed throughthe resin column and eluted with 100 ml 0.01 M nitric acid, and the eluatewas discarded. Trace elements were collected from the column by eluting with50 ml 4 M nitric acid. The eluate was heated to reduce its volume, transferredto a volumetric flask, and diluted to 10 ml.

A3.5 ml portion in a 4 ml polyethylene vial was irradiated for 5 min. Anotherportion, 3.0 ml in a 3.5 ml silica vial, was irradiated for 3 d. After the shortirradiation, 3 ml of the irradiated solution were transferred into an activity-free vial and submitted to γ -ray spectrometry with a Ge(Li) detector coupledto a 4000-channel analyser. After the long irradiation, the sample was allowedto cool for 3 d, then the surface of the silica ampoule was cleaned with dilutenitric acid and the sealed ampoule was placed in the counter (the backgroundactivity of the ampoule was negligible). Gamma-ray energy and the areas underpeaks were calculated by computer. To determine the half-lives of the nuclidesproduced, the counting was repeated at appropriate intervals.

The major interfering elements such as sodium, potassium, bromine, andchlorine (Figs. 5.24 and 5.25) were completely removed from the column with100 ml 0.01 M nitric acid, whereas many trace elements were quantitativelyretained. These elements were eluted with the succeeding 50 ml of 4 M nitricacid.

The recovery ratios indicate that the added traces of cadmium, cerium, cop-per, lanthanum, manganese, scandium, and zinc are quantitatively recovered.The recoveries of barium, cobalt, bromium, iron, uranium, and vanadium werealso satisfactory.


Figure 5.24. Gamma-ray spectrum of preconcentrated river water after short irradiation.Irradiation time 5 min; thermal-neutron flux 1 ×1013 n cm–3 s–1; decay time 3 d; countingtime 1000 s. Source: [627]

Recoveries of calcium and magnesium were very poor for seawater. Thereasons for this may be connected with matrix effects.

In order to evaluate the precision of this method, replicate analyses werecarried out by Lee et al. [627] using the proposed procedure, for trace elementsin a seawater sample taken from the Kwangyang Bay (Korea). The resultsshowed satisfactory precision ranging from 0.2 µg/l for cadmium to 250 µg/lfor iron.

Lieser et al. [628] studied the application of neutron activation analysis tothe determination of trace elements in seawater, with particular reference tothe limits ofdetectionand reproducibility obtained fordifferent elementswhencomparing various preliminary concentration techniques such as adsorptionon charcoal, cellulose, and quartz, and complexing agents such as dithizoneand sodium diethyldithiocarbamate.

In these procedures 1 litre of seawater was shaken with 60 mg charcoal for15 min. Complexing agents were added in amounts of 1 mg, dissolved in 1 mlof acetone. The pH was 5.5, or it was adjusted to 8.5 by addition of 0.1 Mammonia. The charcoal was filtered off and irradiated. Results of three sets ofexperiments with charcoal alone, charcoal in the presence of dithizone, andcharcoal in the presence of sodium diethyldithiocarbamate are compared. Thefollowing elements are adsorbed to an extent from 75 to 100%: silver, gold,cerium, cadmium, cobalt, chromium, europium, iron, mercury, lanthanum,scandium, uranium, and zinc. The amount of sodium is reduced to about 10–6,bromine to about 10–5, and calcium to about 10–2.


Figure 5.25. Gamma-ray spectrum of preconcentrated seawater after long irradiation. Irra-diation time 3 days; thermal-neutron flux 1 ×1013 n cm–3 s–1; decay time 3 d; counting time1000 s. Source: [627]

5.74.16Isotope Dilution Mass Spectrometry

This technique has been used to determine a number of elements in sea-water, including lithium [826], barium [74], lead [827], rubidium [840], ura-nium [828], and copper [298, 299]. It has not been extensively applied.


The earlier stable isotope dilution mass spectrographic work was accom-plished with a thermal ion mass spectrometer which had been specifically de-signed for isotopeabundancemeasurements.However, Leipziger [829]demon-strated that the spark source mass spectrometer could also be used satisfac-torily for this purpose. Although it did not possess the excellent precision ofthe thermal unit, Paulsen and coworkers [830] pointed out that it did havea number of important advantages.

In the analysis of seawater, isotope dilution mass spectrometry offers a moreaccurate and precise determination than is potentially available with other con-ventional techniques such as flameless AAS or ASV. Instead of using externalstandards measured in separate experiments, an internal standard, which isan isotopically enriched form of the same element, is added to the sample.Hence, only a ratio of the spike to the common element need be measured. Thequantitative recovery necessary for the flameless atomic absorption and ASVtechniques is not critical to the isotope dilution approach. This factor can be-come quite variable in the extraction of trace metals from the salt-laden matrixof seawater. Yield may be isotopically determined by the same experiment or bythe addition of a second isotopic spike after the extraction has been completed.

An outline of the elements in seawater that can be analysed by isotopedilution techniques has been presented by Chow [831]. Most of the subsequentwork concerns the analysis of lead in seawater [184, 831–835]. The extensionof the technique to the analysis of copper, cadmium, thallium, and lead hasbeen made by Murozumi [835] using a thermal source mass spectrometer,whereas Russell and Sturgeon [937] used a spark source mass spectrometer inthe determination of iron, cadmium, zinc, copper, nickel, lead, and uranium inseawater. It may be noted that thermal source mass spectrometry has two primeadvantages over spark source mass spectrometry that may be of particularvalue in seawater analysis. One is the much higher precision (about 0.1%)available from the instrument, a factor important in assessing the stabilityof the trace metals in seawater with time or the extraction technique itself.Secondly, the ratios to be determined may be measured at various filamentcurrents (temperature) for an internal corroboration. This is of particularimportance in detecting isobaric interferences that might lead to spuriousratios and hence misleading results.

One of the advantages of the isotope dilution technique is that the quantita-tive recovery of the analytes is not required. Since it is only their isotope ratiosthat are being measured, it is necessary only to recover sufficient analyte tomake an adequate measurement. Therefore, when this technique is used in con-junction with graphite furnace atomic absorption spectrometry, it is possibletodetermine the efficiencyof thepreconcentration step.This is particularly im-portant in the analysis of seawater, where the recovery is very difficult to deter-mine by other techniques, since the concentration of the unrecovered analyte isso low. In using this technique, one must assume that isotopic equilibrium hasbeen achieved with the analyte, regardless of the species in which it may exist.


The thermal ion mass spectrometer was specifically developed for the mea-surement of isotope abundances and is capable of excellent precision. Althoughthe spark source mass spectrometer used in this work lacks some of this preci-sion, it has proved very useful in stable isotope dilution work. It has a numberof advantages, including greater versatility, relatively uniform sensitivity, andbetter applicability to a wide range of elements.

Heavy Metals

Stuckas and Wong [836] have investigated the feasibility of using a thermalsource mass spectrometer in the isotope dilution analysis of copper, cadmium,lead, zinc, nickel, and iron in seawater. The approach basically follows thatwhich had been successfully employed in the analysis of lead in seawater [834].Great importance was attached to the definition of the blank and the initialclean-up schedule. Once the operating parameters had been established bythe analysis of pure isotopic spikes, the various components that constitutea blank were identified and minimised where possible using ultraclean-roomtechniques. The subsequent extractions of reagent water and seawater by dithi-zone and by ammonium pyrrolidine–diethyldithiocarbamate ion exchangeresins were evaluated for suitability for the isotope dilution mass spectromet-ric approach, yield, and contamination levels. The results obtained on seawaterwere corroborated by two independent techniques, ASV and flameless AAS.Satisfactory agreement was obtained in most cases.

Wu and Boyle [837] have developed a method using magnesium hydroxidecoprecipitation and isotopic dilution mass spectrometry to determine lead,copper, and cadmium in 1 ml seawater samples, with detection limits of 1, 40,and 5 pM, respectively.

Heavy Metals, Uranium

Mykytiuk et al. [184] have described a stable isotope dilution spark source massspectrometric method for the determination of cadmium, zinc, copper, nickel,lead, uranium, and iron in seawater, and have compared results with thoseobtained by graphite furnace atomic absorption spectrometry and inductivelycoupled plasma emission spectrometry. These workers found that to achievethe required sensitivity it was necessary to preconcentrate elements in theseawater using Chelex 100 [121] followed by evaporation of the desorbedmetal concentrate onto a graphite or silver electrode for isotope dilution massspectrometry.

Caesium and Chromium

Riel [838] extractedceasiumandchromiumfromseawaterusinganunderwaterprobe and X-ray spectrometry unit.


Lanthanides

See Sect. 5.49.15.


Aluminium, Iron, and Manganese

Nagaosa et al. [839] simultaneously separated and determined these elementsin seawater by high-performance liquid chromatography (HPLC) using spec-trophotometric and electrochemical detectors.

Copper, Nickel, and Vanadium

In a method for the determination of copper, nickel, and vanadium in sea-water, Shijo et al. [840] formed complexes with 2-(5-bromo-2 pyridylazo)-5-(N-propyl-N-sulfopropylamino) phenol and extracted these from the seawaterwith a xylene solution of capriquat. Following back-extraction into aqueoussodium perchlorate, the three metals were separated on a C18 column by HPLCusing a spectrophotometric detector.

Transition Metals

Riccardo et al. [841] showed that chitosan is promising as a chromatographiccolumn for collecting traces of transition elements from salt solution andseawater, and for recovery of trace metal ions for analytical purposes. Tracesof transition elements can be separated from sodium and magnesium, whichare not retained by the chitosan.

5.74.18Metal Speciation

Examining the vertical distribution of trace elements in the ocean yields valu-able clues about their chemical behaviour and cycling mechanisms in the watercolumn [263,519, 842]. The next step in refining our understanding of the ele-ments’ marine chemistry is to examine their chemical speciation. Knowledgeof the chemical forms of trace elements in seawater, and their relative reac-tivity and distribution coefficients for uptake on particles, greatly assists theinterpretations of observed vertical distributions or removal rates. This infor-mation is particularly important for estuarine or coastal environments where,because of rapidly changing conditions of the water column, steady-state ver-tical distributions of elements cannot be used to understand their reactivities.Furthermore, these waters receive large natural or anthropogenic inputs of


potentially harmful substances. With their often restricted circulation, it isimperative to know what controls the movement of these chemicals throughthe system, in order to predict whether they will be concentrated locally, andin what form, or if they will be diluted into the ocean [843, 844].

Chemical separation studies can help to understand the fate of various traceelements by addressing the following questions:

1. Do the elements associate rapidly with a particulate phase and settle to thebottom to accumulate, or are they transported with bottom sediments?

2. To what extent, and at what rate, do they become involved in the biologicalcycle, either(a) through complexation with dissolved organic matter, which may keep

them in solution, or(b) by bioaccumulation in organisms and subsequent transport and recy-

cling with biotic components?3. Do the trace elements remain in true solution, to be diluted throughout the

ocean by mixing?

To begin to answer these questions, the chemical forms and removal rates ofmore than 20 trace metals and metalloids, added in radioactive form to largesynthetic ecosystems (“microcosms”), have been studied.

Chemical speciation studies of trace metals in seawater are of two types. Onegroup consists of theoretical equilibrium models based on known thermody-namic data, derived generally from studies of less complex solutions [843,845–847]. The second approach (Amdurer [847]) is to determine directly the chem-ical forms of trace elements by electrochemical techniques or wet-chemicalseparation schemes [276,848,851]. The problem with this approach is that the“species” identified are often operationally defined by the method used, mak-ing it difficult to compare the results of various methods or to compare theseresults to equilibrium predictions. Nevertheless, these “operational” systemsmay be more useful for predictive purposes. The use of radiotracers made itpossible to simultaneously examine the relative behaviour of a large numberof trace elements with widely-differing chemical properties. By studying thechanges in chemical form with time following addition of the tracers, it is pos-sible to infer the reactivity, chemical transformations in the water column, andmajor removal mechanisms of these elements. These studies provide unique“fingerprints” of each trace element according to the phases in which it occurs.This approach permits the grouping of elements with similar chemical formsand removal patterns. If we understand the pathways of an element througha particular environment, we can then predict the behaviour of like elementsin similar systems.

The experiments performed by Amdurer [847] were not intended to studychemical speciation or cycling of natural elements in the ecosystem. To dothis, the tracers must be fully equilibrated with all of the reactive (i.e., non-matrix) phases of the stable elements. This equilibration may require a time


period equal to or greater than the duration of the experiment. There is ampleevidence that 65Zn, for example, may take years to fully equilibrate with thenatural organically-complexed fraction of stable zinc (II) [855], and the samemay be true of 55Fe [850]. Rather, the radiotracers, added in ionic form, areanalogs for the low-level input of similarly dissolved pollutant metals, such asmay occur in some industrial or municipal effluents.

Van den Berg [852] has studied trace metal speciation in seawater. Thisconcentrates on the complex-forming transition metals, copper, lead, zinc, andcadmium. Both inorganic and organic speciation interactions are discussed.Studiesofboth typesof species in seawaterare reviewed.Thevaluesdeterminedby these workers have been used to calculate the products of stability constantsand ligand concentrations in seawater as a measure of the speciation of themetal ions. The results of a number of recent studies of interactions betweenmetal ions and organic matter in seawater were compared. Organic–metalinteractions can be considerable, at least in surface waters. It is not yet knownwhether deep sea samples are similarly affected but, since the compounds areprobably derived from primary production, the highest ligand concentrationsare to be expected in the surface waters.

Stolzberg [143] has discussed potential inaccuracies in trace metal specia-tion measurement in the determination of copper and cadmium by differentialpulse polarography and ASV.

This author reviews the limitations of ASV and differential pulse polarog-raphy in determining trace metal speciations, and thereby bioavailability andtransportpropertiesof tracemetals innaturalwaters. Inparticular, it is stressedthat nonuniform distribution of metal–ligand species within the polarographiccell represents another limitation inherent in electrochemical measurement ofspeciation. Examples relate to the differential pulse polarographic behaviourof cadmium complexes of NTA and EDTA in seawater.

Midorikawa et al. [853] have conducted studies on the complexing ability ofnatural ligands inseawater forvariousmetal ionsusing ion-selectiveelectrodes.Organic ligands (nominal molecular weight around 1000) are concentratedfrom seawater and desalted by lycophylisation and dialysis. The concentratedsolution of natural ligands is electrodialysed with ethylene diamine tetra-aceticacid to remove metal ions. The demetallised ligands obtained in this way aretitrated with a metal ion to determine the complexing ability of the naturalligands. The advantages of this method are as follows: the formation of a com-plex between organic ligands and a specific metal ion can be studied withoutconsideration of simultaneous side reactions; samples from distinct sources(saline and fresh water, biological, sedimentary, etc.), can be compared on thesame basis; repeated measurements can be made with the same sample afterremoval of the exogenously added metal ions. The ability of natural ligands inseawater to form complexes with copper (II) and cadmium (II) is discussed.

Mackey [854] studied the suitability of Amberlite XAD-1 resin for extractingorganic complexes of copper, zinc, and iron from seawater. The results suggest


Tabl

e5.

12.

Pre-

conc

entr

atio

nof

cati

ons

Cat

ion

Pre-

conc

entr

atio

nm

etho

dA

naly

tica

lfini

shD

etec

tion

limit

Ref

eren

ce

Alu

min

ium

Flow

inje

ctio

nw

ith

in-l

ine

pre-

conc

entr

atio

non

colu

mn

ofre

sin

imm

obili

sed-

8-hy

drox

yqui

nolin

eSp

ectr

ofluo

rim

etry

–[8

59]

Alu

min

ium

Ext

ract

ion

wit

htr

ifluo

ro-a

cety

lace

tone

into

luen

eA

naly

sis

ofto

luen

eex

trac

tby

GC

6pg

Ali

n2

µle

xtra

ct[1

8]A

lum

iniu

mEx

trac

tion

ofpy

roca

tech

olvi

olet

com

plex

wit

hch

loro

form

Spec

trop

hoto

met

ryat

590

nm<

0.1

µg/

l[2

6]A

ntim

ony

Sb(I

II)

and

Sb(V

I)ad

sorb

edas

amm

oniu

mpy

rrol

idin

edi

thio

carb

amat

eco

mpl

exes

onto

C18

bond

edsi

lica

Gra

phit

efu

rnac

eA

AS

0.05

µg/

l[8

60]

Ant

imon

y12

5 Sb

adso

rbed

onto

MnO

2γ

-Ray

spec

trom

etry

0.4

mB

q/l

[587

]A

ntim

ony

Sb(I

II)

and

Sb(V

)co

nver

ted

tohy

drid

es,c

old

trap

ping

inA

AS

0.04

µg/

l[5

9,60

,63,

tolu

ene

861]

Ant

imon

ySb

(III

)an

dSb

(V)

adso

rbed

onam

mon

ium

pyrr

olid

ine

dith

ioca

rbam

ates

onC

18bo

nded

silic

aG

raph

ite

furn

ace

AA

S0.

05ng

/l[8

60]

Ars

enic

As(

III)

and

As(

V)c

ompl

exed

wit

ham

mom

ium

pyrr

olid

ine

diet

hyld

ithi

oca

rbam

ate

and

extr

acte

dw

ith

chlo

rofo

rmC

hlor

ofor

mex

trac

tex

amin

edby

NA

A–

[862

]

Ars

enic

Ars

enic

dete

rmin

edby

cath

odic

stri

ppin

gC

SV0.

3nm

ol[6

21]

Bar

ium

Bar

ium

pre-

conc

entr

ated

onca

tion

exch

ange

resi

n,th

enex

trac

tion

wit

hni

tric

acid

Gra

phit

efu

rnac

eA

AS

–[7

8]

Ber

ylliu

mB

eryl

lium

pre-

conc

entr

ated

asac

etyl

acet

one

com

plex

onto

acti

vate

dch

arco

alC

harc

oali

ntro

duce

ddi

rect

lyin

togr

aphi

tefu

rnac

eA

AS

0.6

ng/l

0.00

06µ

g/l

[79]

Bis

mut

hLi

quid

–liq

uid

extr

acti

onin

toxy

lene

asth

eam

mon

ium

pyrr

olid

ine

dith

ioca

rbam

ate

com

plex

Ele

ctro

ther

mal

AA

S0.

3pp

tor

0.00

05µ

g/l

[95]

Cad

miu

mE

xtra

ctio

nof

amm

oniu

mpy

rrol

idin

edi

thio

carb

amat

eco

mpl

exw

ith

chlo

rofo

rmA

AS

0.00

06µ

g/l

[134

,863

]

Cad

miu

mC

o-pr

ecip

itat

ion

wit

hzi

rcon

ium

hydr

oxid

e,di

ssol

utio

nof

prec

ipit

ate

inhy

droc

hlor

icac

idSq

uare

wav

epo

laro

grap

hyat

–0.6

V–

[144

]

Cad

miu

mC

adm

ium

colle

cted

onC

hele

x-10

0re

sin

and

elut

ion

wit

hni

tric

acid

Gra

phit

efu

rnac

eA

AS

[118

]


Tabl

e5.

12.

(con

tinu

ed)

Cat

ion

Pre-

conc

entr

atio

nm

etho

dA

naly

tica

lfini

shD

etec

tion

limit

Ref

eren

ce

Cae

sium

Che

lati

onas

copp

erhe

xacy

ano

ferr

ate(

II)

onsi

lica

AA

S–

[864

]C

aesi

um13

7 Cs

adso

rpti

onon

stro

ngca

tion

exch

ange

resi

nam

mo-

nium

hexa

cyan

oco

balt

ferr

ate,

diss

olut

ion

inhy

droc

hlor

ic– hy

drofl

uori

cac

id

βco

unte

dfo

r13

7 Cs

10pC

i/l

[865

]

Cae

sium

Csc

ompl

exed

wit

h4-

tetr

abut

yl-2

(α-m

ethy

lben

zyl)

phen

olan

dso

lven

text

ract

edA

AS

–[1

49]

Cal

cium

Ads

orpt

ion

onC

hele

x-10

0re

sin,

deso

rbed

wit

hhy

droc

hlo-

ric

acid

Flam

eph

otom

etry

at62

2nm

–[1

67]

Cer

ium

Ce(

IV)

co-p

reci

pita

tion

wit

hfe

rric

hydr

oxid

e,di

ssol

utio

nin

hydr

ochl

oric

acid

,the

npa

ssed

thro

ugh

aco

lum

nof

bis

(2et

hyl

hexy

l)ph

osph

ate

onpo

ly(v

inyl

chlo

ride

),el

uted

wit

h0.

3M

perc

hlor

icac

id

Spec

trofl

uori

met

ryat

350

nm(e

xcit

atio

n25

5nm

)–

[626

]

Chr

omiu

mC

o-pr

ecip

itat

ion

ofC

r(IV

)an

dC

r(II

I)w

ith

lead

sulfa

teN

AA

0.1

µg/

l(8

00µ

lsam

ple)

[95]

Chr

omiu

mIm

mob

ilise

ddi

phen

ylca

rbaz

one

Zee

man

AA

S–

[178

]C

hrom

ium

Cr(

III)

conv

erte

dto

Tris

(1,1

,1,t

riflu

oro

2,4

pent

aned

iono

)Is

otop

edi

luti

onG

C-M

S[1

81,1

83,

chro

miu

m(I

II),

extr

acte

dw

ith

hexa

ne86

7]C

hrom

ium

Cr(

III)

adso

rbe

onsi

lica,

oxid

ised

toch

rom

ium

(IV

)by

UV

dige

stio

nC

SV0.

0016

µg/

l[1

86]

Cob

alt

Com

plex

atio

nw

ith

2-et

hyl-

amin

o-5-

nitr

osop

heno

l,ex

-tr

acti

onw

ith

dich

loro

etha

neSp

ectr

opho

tom

etry

at40

2nm

<0.

07µ

g/l

[224

]

Cob

alt

Pre-

conc

entr

atio

non

man

gane

sedi

oxid

e,so

luti

onin

hy-

droc

hlor

icac

idPu

lse

pola

rogr

aphy

–[2

31]

Cob

alt

Com

plex

atio

nw

ith

4-(2

-thi

azol

ylaz

o)re

sorc

inol

,ads

orp-

tion

onX

AD

-4,r

esin

elut

ion

wit

hm

etha

nol:c

hlor

ofor

mG

raph

ite

furn

ace

AA

S–

[234

]

Cob

alt

Co-

prec

ipit

atio

nw

ith

sodi

umdi

ethy

ldit

hio

carb

amat

eX

-ray

fluor

esce

nce

spec

trom

etry

0.3

µg/

l[5

11]


Tabl

e5.

12.

(con

tinu

ed)

Cat

ion

Pre-

conc

entr

atio

nm

etho

dA

naly

tica

lfini

shD

etec

tion

limit

Ref

eren

ce

Cob

alt

Co(

III)

adso

rbed

onC

18bo

nded

silic

aLa

ser

exci

ted

atom

icflu

ores

cenc

esp

ectr

omet

ry–

[241

]

Cob

alt

Co(

III)

adso

rbed

on8-

quin

olol

adso

rbed

onsi

lica

Che

mic

allu

min

esce

nce

–[2

44]

Cop

per

Ads

orpt

ion

onA

mbe

rlit

eX

AD

-4M

isce

llane

ous

–[8

54]

Cop

per

Ext

ract

ion

ofam

mon

ium

pyrr

olid

ine

dith

ioca

rbam

ate

com

plex

wit

hM

IBK

Mis

cella

neou

s<

0.5

µg/

l[3

74]

Cop

per

Ads

orpt

ion

onch

itos

an,e

xtra

ctio

nw

ith

1,10

phen

anth

ro-

line

AA

S<

0.1

µg/

l[8

41]

Cop

per

Extr

acti

onw

ith

8-qu

inol

olin

ethy

lpro

pion

ate

Ele

ctro

nsp

inre

sona

nce

spec

tros

copy

<2

µg/

l[3

00]

Cop

per

Ads

orpt

ion

onhy

drox

yqui

nolin

e,di

thiz

one

ordi

ethy

ldit

hioc

arba

mat

em

odifi

edch

arco

alM

isce

llane

ous

–[3

01]

Cop

per

Ads

orpt

ion

onsi

lica

mod

ified

wit

h8-

quin

olol

Flow

inje

ctio

nA

AS

–[8

68]

Cop

per

Ads

orpt

ion

onC

18bo

nded

silic

a–

–[3

32]

Cop

per

Ads

orpt

ion

onba

sic

Am

berl

ite

GC

-400

anio

nex

chan

gere

sin,

elut

ion

wit

hni

tric

acid

AA

S0.

013

µg/

l[8

69]

Cop

per

Ads

orpt

ion

onC

18re

sin

Lase

rex

cite

dat

omic

fluor

esce

nce

spec

trom

etry

wit

ha

grap

hite

elec

trot

herm

alat

omis

er

0.00

1µ

g/l

[241

]

Indi

umC

o-pr

ecip

itat

ion

wit

hga

llium

phos

phat

eE

lect

roth

erm

alA

AS

0.3

µg/

l(5

00m

lsam

ple)

[870

]

Iron

Con

vers

ion

toba

thop

hena

nthr

olin

eco

mpl

exad

sorp

tion

oman

ion

exch

ange

resi

n(A

mbe

rlit

eA

27)

Den

sito

met

ric

scan

ning

ofre

sin

at55

0nm

–[3

50]

Iron

Iron

co-p

reci

pita

ted

wit

hte

trap

heny

lbo

ron/

1,10

-ph

enan

thro

line,

prec

ipit

ate

diss

olve

din

amm

onia

cal

etha

nol

Mis

cella

neou

s–

[568

]


Tabl

e5.

12.

(con

tinu

ed)

Cat

ion

Pre-

conc

entr

atio

nm

etho

dA

naly

tica

lfini

shD

etec

tion

limit

Ref

eren

ce

Iron

Ads

orbe

don

ferr

ozin

eim

preg

nate

dC

18re

sin

Spec

trop

hoto

met

ry0.

1m

M/l

[351

]Ir

onPr

e-co

ncen

trat

edon

8-qu

inol

olch

elat

ing

resi

nC

hem

ilum

ines

cenc

e50

pmol

/l0.

0025

µg/

l(1

8m

lsam

ple)

[349

]

Lead

Ele

ctro

chem

ical

pre-

conc

entr

atio

nof

cont

rolle

dpo

tent

ials

Ele

ctro

chem

ical

pre-

conc

entr

atio

nA

AS

–[3

65,3

95]

Lead

Lead

extr

acte

dfr

omw

ater

wit

hch

loro

form

icdi

thiz

one

Gra

phit

efu

rnac

eA

AS

–[3

64]

Lead

Lead

co-p

reci

pita

ted

wit

hN

i 3K

2[F

e(C

N) 6

]X

-ray

fluor

esce

nce

spec

trom

etry

2–

4µ

gab

solu

te[5

41]

Lead

Lead

pre-

conc

entr

ated

onC

hele

x-10

0re

sin

Indu

ctiv

ely

coup

led

plas

ma

atom

icem

issi

onsp

ectr

omet

ry0.

6ng

/l0.

0006

µg/

lco

mpa

red

to60

ng/l

(0.0

6µ

g/l)

bydi

rect

nebu

lisat

ion

[871

,872

]

Lead

Com

plex

atio

nw

ith

amm

oniu

mpy

rrol

idin

edi

thio

carb

a-m

ate,

colle

ctio

non

C18

mic

roco

lum

nG

raph

ite

furn

ace

AA

S3.

5ng

/l[8

73]

Lead

Com

plex

atio

nw

ith

DD

PAad

sorp

tion

onC

18m

icro

colu

mn,

elut

ion

wit

hm

etha

nol

Lase

ren

hanc

edio

nisa

tion

3.3

µg/

l(1

5m

lsam

ple)

[360

]

Man

gane

seM

anga

nese

com

plex

edw

ith

amm

oniu

man

ddi

ethy

lam

-m

oniu

mdi

ethy

ldit

hioc

arba

mat

ean

dex

trac

ted

wit

hFr

eon

Ele

ctro

ther

mal

AA

Sof

Freo

nex

trac

t–

[65,

448]

Man

gane

seEx

trac

tion

wit

hch

loro

form

ic8-

quin

olol

,bac

kex

trac

tion

into

aque

ous

nitr

icac

idG

raph

ite

furn

ace

AA

S<

30ng

/l0.

03µ

g/l

[432

,444

]

Man

gane

seEx

trac

tion

ofm

anga

nese

wit

ha

hept

ane

solu

tion

ofbi

s(2

-eth

ylhe

xyl)

phos

phat

eSp

ectr

opho

tom

etri

c–

[874

]

Man

gane

seA

dsor

ptio

non

Che

lex-

100,

extr

acti

onof

man

gane

sew

ith

Gra

phit

efu

rnac

eA

AS

–[1

29,4

50,

2N

nitr

icac

id45

1]M

anga

nese

Com

plex

atio

nw

ith

theo

nyl-

trifl

uoro

acet

one

and

dibe

nzol

-8-

crow

n-6

AA

S0.

1µ

g/l

[875

]


Tabl

e5.

12.

(con

tinu

ed)

Cat

ion

Pre-

conc

entr

atio

nm

etho

dA

naly

tica

lfini

shD

etec

tion

limit

Ref

eren

ce

Mer

cury

Hg

redu

ced

toel

emen

talm

ercu

ry,e

xtra

ctw

hich

issw

ept

AA

S0.

2pM

(abs

olut

e)[4

60,4

67,

onto

gold

foil

469,

472,

478,

481,

875–

877]

Mer

cury

Mer

cury

conv

erte

dto

diet

hyld

ithi

ocar

bam

ate,

conc

en-

trat

edX

AD

-2re

sin,

elut

edw

ith

met

hano

lichy

droc

hlor

icac

id

Flam

eles

sA

AS

0.1

ng/l

[878

]

Mer

cury

Mer

cury

com

plex

edw

ith

amm

oniu

mte

tra

met

hyle

ne-

dith

ioca

rbam

ate,

extr

acte

dw

ith

chlo

rofo

rmC

hlor

ofor

mex

trac

tan

alys

edby

grap

hite

tube

AA

S<

5ng

/lH

gab

solu

te[4

79,8

79]

Mer

cury

Ano

dic

stri

ppin

gvo

ltam

met

ryw

ith

glas

syca

rbon

elec

-tr

ode

spin

coat

edw

ith

4,7,

13,1

6,21

,42

hexa

oxa,

1,10

diaz

-ab

icyc

lo[8

,8,8

]he

xaco

sane

––

[483

]

Mer

cury

Com

pari

sion

ondi

ffer

ent

mer

cury

chel

atin

gag

ents

ad-

sorb

edon

tosi

lica

C18

resi

nC

old

vapo

urA

AS

0.01

6µ

g/l

[880

]

Mol

ybde

num

Co-

prec

ipit

atio

nm

etho

ds–

–[8

81–8

85]

Mol

ybde

num

Solv

ente

xtra

tcti

onm

etho

ds–

–[2

1,51

4,88

6,88

8]M

olyb

denu

mC

o-cr

ysta

llisa

tion

met

hods

––

[21,

514,

886–

888]

Mol

ybde

num

Sorp

tion

onso

lids

orC

hito

san

mod

ified

cellu

lose

AA

S–

[137

,502

,50

3,88

8,88

9]M

olyb

denu

mC

atio

nex

chan

gead

sorp

tion

––

[891

]M

olyb

denu

mC

atio

nex

chan

gead

sorp

tion

onC

hele

x-10

0–

–[8

92,8

93]

Mol

ybde

num

Ani

onex

chan

gere

sin

adso

rpti

on(B

iora

dA

g1-X

8),e

luti

onw

ith

amm

onia

Gra

phit

efu

rnac

eA

AS

<10

µg/

l[5

08]


Tabl

e5.

12.

(con

tinu

ed)

Cat

ion

Pre-

conc

entr

atio

nm

etho

dA

naly

tica

lfini

shD

etec

tion

limit

Ref

eren

ce

Mol

ybde

num

Ani

on-e

xcha

nge

onD

owex

1-X

8re

sin

orA

mbe

rlit

eG

C-4

0Sp

ectr

opho

tom

etry

–[4

97–5

00,

resi

n51

6,89

4]M

olyb

denu

mA

dsor

ptio

nof

mol

ybde

num

pyrr

olid

ine

dith

ioca

rbam

ate

onch

arco

al–

–[8

95]

Mol

ybde

num

Co-

prec

ipit

atio

nw

ith

Zr(

OH

) 4In

duct

ivel

yco

uple

dpl

asm

aA

ES–

[896

]N

icke

lN

icke

lis

co-p

reci

pita

edw

ith

sodi

umdi

ethy

ldit

hioc

arba

-m

ate

and

the

prec

ipit

ate

diss

olve

din

nitr

icac

idA

AS

0.5

ng/l

[897

]

Nic

kel

Nic

kel

adso

rbed

onto

poly

(tri

amin

ophe

nyl)

glyo

xal,

des-

orbe

dw

ith

amm

oniu

mpy

rrol

idin

edi

thio

carb

amat

ein

MIK

B

AA

S–

[897

]

Nic

kel

Solv

ente

xtra

ctio

nte

chni

ques

Mis

cella

neou

s–

[122

]N

icke

lC

o-pr

ecip

itat

ion

tech

niqu

esM

isce

llane

ous

–[5

18–5

20]

Nic

kel

Nic

kel

dim

ethy

lgl

yoxi

me

com

plex

extr

acte

dfr

omw

ater

wit

hch

loro

form

AA

S5

µg/

l0.

0005

µg/

l[8

99]

Nic

kel

Nic

kelc

onve

rted

toN

i(C

O) 4

byso

dium

boro

hydr

ide.

Col

dtr

appe

dus

ing

liqui

dni

trog

enA

AS

wit

hqu

artz

tube

burn

erSu

b-ng

[524

]

Palla

dium

Ads

orbe

don

toliq

uid

mem

bran

eor

tri-

N-o

ctyl

-am

ine

––

[530

]Po

loni

umPo

loni

umel

ecro

depo

site

dfr

omw

ater

onto

carb

onro

d,st

ripp

edfr

omro

dsan

dau

topl

ated

onto

silv

erco

unti

ngdi

scs

Aut

opla

ting

208 P

uco

unte

d–

[901

,902

]

Plut

oniu

mPu

co-p

reci

pita

ted

wit

hfe

rric

hydr

oxid

e,pr

ecip

itat

edi

s-so

lved

inac

idA

nion

exch

ange

chro

mat

ogra

phy

–[9

03,9

04]

Plut

oniu

mPl

uton

ium

(IV

)co

nver

ted

toxy

leno

lor

ange

com

plex

onX

AD

-2re

sin

α-r

aysp

ectr

omet

ryof

resi

n–

[905

]

Rhe

nium

Ads

orpt

ion

onD

eaci

dite

FFan

ion

exch

ange

resi

n,el

utio

nof

resi

nw

ith

4M

nitr

icac

idN

eutr

onac

tiva

tion

anal

ysis

0.06

ng/l

(15

mls

ampl

e)[5

39]


Tabl

e5.

12.

(con

tinu

ed)

Cat

ion

Pre-

conc

entr

atio

nm

etho

dA

naly

tica

lfini

shD

etec

tion

limit

Ref

eren

ce

Rub

idiu

mR

ubid

ium

co-p

reci

pita

ted

wit

hN

i 3K

2[Fe

(CN

) 6] 2

X-r

ayflu

ores

cenc

esp

ectr

omet

ry2

µg/

labs

olut

e[5

41]

Sele

nium

Sele

nium

conv

erte

dto

3,3-

diam

inob

enzi

dine

piez

sele

nol,

whi

chis

extr

acte

din

toto

luen

ean

dba

ck-e

xtra

cted

into

0.5

Mhy

droc

hlor

icac

id

Com

plex

depo

site

don

am

ercu

ryel

ectr

ode

at–0

.45

V0.

01µ

gSe

abso

lute

[542

]

Sele

nium

Sele

nium

(IV

)am

mon

ium

pyrr

olid

ine

diet

hyld

ithi

ocar

ba-

mat

eco

mpl

exad

sorb

edon

toC

18bo

nded

silic

a,th

ende

s-or

bed

Gra

phit

efu

rnac

eA

AS

7ng

/l[8

60]

Sele

nium

Con

vers

ion

ofse

leni

umin

tose

leni

umhy

drid

eSe

leni

umhy

drid

esw

ept

into

quar

tztu

beat

400

◦ C,a

naly

sed

byhy

drid

ege

nera

tion

Gra

phit

efu

r-na

ceA

AS

20ng

/l[9

06]

Sele

nium

Sele

nium

co-p

reci

pita

ted

wit

hFe

(OH

) 3,

prec

ipit

ate

dis-

solv

edin

hydr

ochl

oric

acid

,co

nver

ted

to5-

nitr

o-pi

azse

leno

l

Gas

chro

mat

ogra

phy

wit

hel

ec-

tron

capt

ure

dete

ctor

5ng

/l[9

07]

Sele

nium

Sele

nium

com

plex

edw

ith

4-ni

tro-

o-ph

enyl

ened

iam

ine,

extr

acte

din

toto

luen

eG

asch

rom

atog

raph

y–

[555

]

Sele

nium

Ads

orpt

ion

collo

idflo

tati

onSp

ectr

opho

tom

etry

ofm

ethy

lene

blue

com

plex

–[5

17,5

46]

Silv

erSi

lver

pre-

conc

entr

ated

onD

eaci

dite

FFIP

anio

nex

chan

gere

sin,

elut

edw

ith

thio

urea

Neu

tron

acti

vati

onan

alys

is<

40ng

/l[5

59]

Silv

erSi

lver

com

plex

edw

ith

amm

oniu

mdi

thio

phos

phor

ate

and

o,o

diet

hyle

ster

then

adso

rbed

onto

carb

on,s

ilver

deso

rbed

wit

hni

tric

acid

Ele

ctro

ther

mal

AA

S0.

3ng

/l0.

0003

µg/

l[9

08]

Stro

ntiu

mSt

ront

ium

com

plex

edw

ith

1-ph

enyl

-3-m

ethy

l-4-

benz

oyl-

5-py

razo

neA

AS

0.00

7ng

/l[9

09]

Tech

neti

umPr

e-co

ncen

trat

edon

anio

nex

chan

geco

lum

n–

3m

Bq/

l[9

10]


Tabl

e5.

12.

(con

tinu

ed)

Cat

ion

Pre-

conc

entr

atio

nm

etho

dA

naly

tica

lfini

shD

etec

tion

limit

Ref

eren

ce

Tech

neti

umC

o-pr

ecip

itat

ion

wit

hFe

(OH

) 3In

duct

ivel

yco

uple

dpl

asm

am

ass

spec

trom

etry

–[9

11]

Tellu

rium

Tellu

rium

copr

ecip

itat

edw

ith

Mg(

OH

) 2A

AS

0.5

pmol

/l[8

63]

Tin

Tin

com

plex

edw

ith

sodi

umdi

ethy

ldit

hio

carb

amat

eSp

ectr

opho

tom

etry

–[5

88]

Tin

Tin

adso

rbed

onto

Dow

ex1-

X8

anio

nex

chan

geco

lum

n,de

sorb

edw

ith

2M

nitr

icac

idSp

ectr

opho

tom

etry

wit

hca

tech

olvi

olet

–[9

12]

Ura

nium

Liqu

id–l

iqui

dex

trac

tion

ofur

aniu

mN

-phe

nyl-

3-st

yryl

acry

lohy

drox

amic

acid

com

plex

Spec

trop

hoto

met

ry1

mg/

l[5

90]

Ura

nium

Form

atio

nof

trio

ctyl

phos

phin

eco

mpl

ex,e

ther

extr

acti

onFl

uori

met

ric

and

spec

trop

hoto

m-

etry

–[9

13,9

14]

Ura

nium

Chl

orof

orm

extr

acti

onof

uran

ium

quin

olin

eco

mpl

exE

lect

roan

alyt

ical

–[2

91]

Ura

nium

Ura

nium

adso

rbed

asaz

ide

onba

sic

ion

exch

ange

colu

mn,

uran

ium

deso

rbed

wit

h1

Mhy

droc

hlor

icac

idSp

ectr

opho

tom

etry

–[9

15]

Ura

nium

Ura

nium

adso

rbed

onbi

smut

hol(

II)

mod

ified

anio

nex

-ch

ange

resi

n,de

sorb

edw

ith

0.1

Mcy

stei

neFl

uoro

met

ry–

[916

]

Ura

nium

Ura

nium

,by

squa

rew

ave

adso

rpti

vest

ripp

ing

volta

mm

e-tr

y–

0.1

µm

ol/l

[592

]

Ura

nium

Ura

nium

copr

ecip

itat

edw

ith

alum

iniu

mph

osph

ate,

pre-

cipi

tate

diss

olve

din

nitr

icac

idFi

ssio

ntr

ack

met

hod

–[9

17]

Ura

nium

Ads

orpt

ion

onto

collo

idal

ferr

ichy

drox

ide

Spec

trop

hoto

met

ric

at55

0nm

byR

hoda

min

eB

met

hod

–[9

18]

Van

adiu

mV

anad

ium

thio

cyan

ate

com

plex

adsi

orbe

don

Dow

ex1-

X8

Spec

trop

hoto

met

ric

at54

5nm

by–

[597

,602

,an

ion

exch

ange

resi

n,de

sorb

edw

ith

nitr

icac

id2-

pyri

dyla

zore

sorc

inol

met

hod

orA

AS

919]

Van

adiu

mV

anad

ium

com

plex

edw

ith

tetr

aphe

nyl-

arso

nium

chlo

ride

and

tetr

amet

hyle

nedi

thio

-car

bam

ate,

extr

acte

dw

ith

chlo

-ro

form

Gra

phit

efu

rnac

eA

AS

<0.

5µ

g/l

[598

]


Tabl

e5.

12.

(con

tinu

ed)

Cat

ion

Pre-

conc

entr

atio

nm

etho

dA

naly

tica

lfini

shD

etec

tion

limit

Ref

eren

ce

Van

adiu

mA

bsor

ptiv

est

ripp

ing

volt

amm

etry

–70

pM[6

01]

Van

adiu

mC

o-pr

ecip

itat

ion

wit

hfe

rric

hydr

oxid

e,co

balt

amm

oniu

mpy

rrol

idin

edi

thio

carb

amat

eor

amm

oniu

mpy

rrol

idin

edi

thio

carb

amat

e

Mis

cella

neou

s–

[920

]

Zin

cZ

inc

colle

cted

onm

ethy

lca

pryl

amm

oniu

mch

lori

deco

ated

C18

resi

nG

raph

ite

furn

ace

AA

S2.

4µ

g/l

[617

]

Zin

cZ

inc

adso

rbed

onC

hele

x-10

0re

sin,

deso

rbed

wit

hni

tric

acid

Gra

phit

efu

rnac

eA

AS

0.5

µg/

l[6

18]

Zin

cC

ompl

exat

ion

wit

hp-

tosy

l-8-

amin

oqui

nolin

ean

dad

sorp

-ti

onon

cati

onex

chan

gere

sin

Spec

trofl

uori

met

ry0.

1nM

abso

lute

[609

]

Zin

cFo

rmat

ion

ofzi

ncam

mon

ium

pyrr

olid

ine

dith

ioca

rba-

mat

eco

mpl

exC

atho

dic

stri

ppin

gvo

ltam

met

ry–

[619

]

Zir

coni

umC

o-pr

ecip

itat

ion

wit

hfe

rric

hydr

oxid

eSp

ectr

opho

tom

etry

byal

izar

inR

met

hod

–[6

08,9

22]

Sour

ce:A

utho

r’sow

nfil

es


Table5.13. Pre-concentration of metals in sea water chelation-solvent extraction techniquesfollowed by direct AAS and graphite furnace AAS

Metals Chelating agent Solvent Detection limit∗(µg/l)

Reference

Direct AAS

Mn, Fe, Co,Ni, Zn, Pb,Cu

Hexahydroazepine-1-carbodithioate

Butylacetate MnFeCoNiZnPbCu

0.21.50.60.60.42.60.5

[923]

Fe, Pb, Cd,Co, Ni, Cr,Mn, Zn, Cu

Diethyldithiocarbamate MIBK orxylene

[924]

Fe, Cu Ammoniumpyrrolidinedithiocarbamate

MIBK CuFe

< 1< 1

[925]

Graphite furnace AAS

Cu, Ni, Cd Ammoniumpyrrolidinedithiocarbamate

[518]

Ag, Cd, Cr,Cu, Fe, Ni,Pb, Zn

Ammoniumpyrrolidinedithiocarbamate

MIBK AgCdCrCuFeNiPbZn

0.020.030.050.050.200.100.030.03

[926]

Cu, Cd, Zn,Ni

Diethyldithiocarbamate plusAmmoniumpyrrolidinedithiocarbamate

Chloroform CuCdZnNi

1.00.22.010.0

[122]

Cd, Pb, Ni,Cu, Zn

Ammoniumpyrrolidinedithiocarbamateplus diethyldithiocarbamate

Freon No stated [124]

Cd, Cu, Fe Ammoniumpyrrolidinedithiocarbamateplus diethyldithiocarbamate

Freon Not stated [119]

Cd, Zn, Pb,Cu, Fe, Mn,Co, Cr, Ni

Ammoniumpyrrolidine-N-carbodithioateplus 8-hydroxyquinoline

MIBK FeCuPbCdZn

0.080.100.060.020.34

[121]

Cd, Zn, Pb,Fe, Mn, Cu,Ni, Co, Cr,

Dithiocarbamate MIBK Not stated [736]



Metals Chelating agent Solvent Detection limit∗(µg/l)

Reference

Cd, Co, Cu,Fe, Mn, Ni,Pb, Zn


Chloroform CdCuFeMnNiPbZn

< 0.0001< 0.012< 0.02< 0.004< 0.012< 0.016< 0.08

[879]

Cd, Co, Cu,Fe, Mn, Ni,Pb, Zn


Chloroform CdCuFeMnNiPbZn

0.020.240.240.020.080.041.0

[879]

Mn, Cd Ammoniumpyrrolidinedithiocarbamateand diethylammoniumdithylthiocarbamate

Freon MnCd

0.070.027

[448]

Cd, Cu, Fe,Pb, Ni, Zn

Ammoniumpyrrolidinedithiocarbamateand diethylammoniumdithylthiocarbamate

Freon Not quoted [927]

Bi Ammoniumpyrrolidinedithiocarbamate

Xylene 0.003 [95]

Pb, Cd, Co,Cu, Sn, As,Mo


– – [666]

Cd, Co, Cu,Ni, Pb, Zn

Sodium bis (2-hydroxyethyl)dithiocarbamate

– – [928]

Cu, Bi, Cd,Zn, Pb

Diethyl and dibutyl dithio-phosphate

Carbontetrachlo-ride

CuBiCdZnPb

0.60.50.80.80.5

[667]


that the resin also adsorbs small but significant amounts of inorganic ionsfrom seawater, and is therefore unsuitable for use in studies on the speciationof trace metals.

Latouche et al. [855] have reviewed trace metal speciation in seawater.Baskaran et al. [856] have discussed the rapid extraction and determination

of thorium, lead, and radium species from large volumes of seawater.Magni et al. [857] studied the optimisation of the extraction of metal–

humic acid complexes from marine sediments. Polyarylamide gels have been


Table 5.14. Application of Chelex-100 resin to the pre-concentration of metals in seawaterprior to analysis by graphite atomic spectrometry

Elements Concentrationfactor

Eluent Detection limit(µg/l)

Reference

Cd, Co, Cu, Fe,Mn, Ni, Pb, Zn

100:1 2.5 M nitric acid subnanogram [129]

Cu, Cd, Zn, Ni, 120:1 2 M nitric acid CuCdZnNi

0.0060.0060.0150.015

[122]

Cd, Zn, Pb, Cu, Fe,Mn, Co, Cr, Ni

20:1 2.5 M nitric acid Not stated [121]

Cd, Pb, Ni, Cu, Zn, 100:1 2.5 M nitric acid CdPbNiCuZn

0.010.16–0.280.24–0.680.61.8

[821]

Cr, Cu, Mn – – [929]Cd, Cu, Ni, Pb – Nitric acid (pH 5)

plus 7.5 M ammo-nium acetate

– [124]


Table 5.15. Miscellaneous solids used in the pre-concentration of cations in seawater

Cation Solid adsorbent Analytical finish Detectionlimit (µg/l)

Reference

Mo, V Cellulose phosphate ICPAES – [930]Se(IV), Se(VI) Thiol cotton (subsequent

desorption with nitricand magnesium nitrite)

Cathodic strippingvoltammetry

Se(VI) 0.009Se 0.018

[931]

Cd, Co, Cu, Fe,Mn, Ni, Pb, Zn

Maleic acid/ammoniumhydroxide buffer system

– – [932]

Heavy metals Silica gel X-ray flourescence – [813]


used as in situ probes for the determination of trace metals in sediment porewater [858].

5.74.19Metal Preconcentration

The considerable difficulty of trace element analysis in a high salt matrix suchas seawater, estuarine water, or brine is clearly reflected in the literature. Theextremely high concentrations of the alkali metals, alkaline earth metals, and


Table 5.16. Applications of co-precipitating agents to the pre-concentration of cations inseawater

Cation Solid adsorbent Analyticalfinish

Detection limit(µg/l)

Reference

Cd, Zn, Mn, Co,Cu, Ni, V, Cr, Si,B, Be, Ca, Mg, Li,lanthanides

Magnesium hydroxide ICP-AES < 4(200 ml sample)

[933]

Hg, Pb, Cd Copper sulfide AAS – [685]Cd, Cu, Pb Palladium salts AAS – [664]Miscellaneoustrace metals

Maleic acid/ammoniumhydroxide buffer system

– – [934]

Ag(I), Cd(II),Cr(III), Cu(II),Mn(II), Th(IV),U(VI), Zr(IV)

Lead phosphate Neutronactivationanalysis

AgCdCrCuMnThUZr

0.23.2–3.872.70.07–1.70.3–0.90.007–0.01840

[825]

(100 ml sample)


halogens, combined with the extremely low levels of the transition metals andother elements of interest, make direct analysis by most analytical techniquesdifficult or impossible.

Brief mention has been made, particularly in connection with the induc-tively coupled plasma atomic absorption spectrometric technique, of the needto preconcentrate seawater samples prior to the determination of metals, inorder to achieve adequate detection limits.

A variety of preconcentration procedures has been used, including solventextraction of metal chelates, coprecipitation, chelating ion exchange, adsorp-tion onto other solids such as silica-bonded organic complexing agents, andliquid–liquid extraction.

An ideal method for the preconcentration of trace metals from naturalwaters shouldhave the following characteristics: it should simultaneously allowisolation of the analyte from the matrix and yield an appropriate enrichmentfactor; it should be a simple process, requiring the introduction of few reagentsin order to minimise contamination, hence producing a low sample blank anda correspondingly lower detection limit; and it should produce a final solutionthat is readily matrix-matched with solutions of the analytical calibrationmethod.

Much work has been published on preconcentration techniques; this workis reviewed below.


Table 5.17. Miscellaneous supported chelators used for pre-concentrations in seawater

Element Chelating agent Solid ad-sorbentsupport

Analyticfinish

Detection limit(µg/l)

Reference

Cd, Pb, Zn,Cu, Fe, Mn,Ni,

8-Hydroxyquinoline Silica AAS CoFe

0.240

[935]

(50 ml samples)Cd, Cu, Pb,Zn,

(3-mercapto propyl)trimethyl oxysilane

Silica – – [936]

Ag(I),Au(III),Pd(II)

p-dimethyl-aminobenzyldine-rhodamine

Silica – – [937]

Cu, Ni, Cd 8-quinolinol Silica ICPAES 0.016–0.07 [938]Fe, Co, Zn,Cs, Zr

Zirconiumhexacyanoferrate(II)

Cationexchangeresin

γ -spectro-metry

– [939]

Rare earths Bis (2-ethyl-hexyl)hydrogen phosphateand 2-ethyl dihydro-gen phosphate

C18 silica ICPMS – [650]

Cd, Cu Sodium diethyldithiocarbamate

C18 silica AAS 0.004–0.024 [283]

Hg, Pb Tubular membrane Dithio-carbamate

– – [93]

Cd, Cu, Mn,Ni, Pb, Zn

Chelamine resin – – – [941]

Ga, Ti, In 8-Hydroxyquinolineimmobilised resin

– ICPAES 0.001–0.004 [942]

Fe, Ni, Cu,Cd, Mo, Cr,V

Chelating Silica ICPAES – [943]


Firstly, work on preconcentration methods for the single elements is re-viewed in alphabetical order in Table 5.12.

Information on preconcentration of multicomponent mixtures is reviewedin Tables 5.13 to 5.18.

The benefits imparted by preconcentration to improved sensitivity are illus-trated in the example of lead preconcentration on Chelex 100 resin [871, 872],followed by analysis by ICP–AES. Without preconcentration the best detec-tion limit achievable is 60 ng/l, via direct nebulisation. When the Chelex 100preconcentration step is included, the detection limit improves to 0.6 ng/l, i.e.,100 times better, which is a very important improvement achieved in the analy-sis of seawaters. Examinaton of Table 5.12 reveals that the following metals canbedeterminedwithdetection limits in the1–10 ng/l range:beryllium(0.6 ng/l),


Table 5.18. Applications of dithiocarbamate acid derivatives to the preconcentration ofcations in seawater

Cationsdetermined

Organic coprecipitant Analytical finish Detectionlimit (µg/l)

Reference

Hg, Au, Cu Leaddiethyldithiocarbamate

Neutron activa-tion analysis

[820]

Cd, Co, Cu, Hg,Mn, Th, U, V, Zn

1) (1-(2-thiazolylazo)-2-naphthol)2) Pyrrolidinedithiocar-bamate3) N-nitrosophenyl-hydroxylamine

Neutron activa-tion analysis

µg/l level [944]

Heavy metals Ammonium pyrro-lidinedithiocarbamate

X-rayfluorescence

– [945]


bismuth (0.3 ng/l), cadmium (0.6 ng/l), indium (0.3 ng/l), tellurium (0.07 ng/l),lead (0.6 ng/l), mercury (0.1 ng/l), silver (0.3 ng/l), and nickel (0.5 ng/l). In ev-ery case but two, preconcentration involves the solvent extraction of a metalcomplex of the metal followed by AAS. Without preconcentration, detectionlimits would be 50 to 100 times higher. Elements for which preconcentrationachieves detection limits in the 10–100 ng/l range include cobalt (70 ng/l), iron(30 ng/l), copper (10 ng/l), manganese (30 ng/l), molybdenum (10 000 ng/l), ru-bidium (1000 µg/l), selenium (5 ng/l), uranium (20 µg/l), vanadium (500 ng/l),zinc (500 ng/l), antimony (50 ng/l), and chromium (1.6 ng/l).

5.74.20Miscellaneous

Worsfold et al. [960] have discussed the application of flow injection analy-sis with chemiluminescence detection for the shipboard monitoring of tracemetals.

Achterberg et al. [961] have described an electrochemical monitor for nearreal-time determination of dissolved trace metals in marine waters, and alsoin surface seawater [962].

References

1. Lin B, Wu H, Cai Z (1987) Yuanzineng Kexue Jishu 21:4472. Ishibashi M, Kawai T (1952) Nippon Kagaku Zasshi 73:3803. Ishibashi M, Motojima K (1952) Nippon Kagaku Zasshi 73:4914. Motojimi K (1955) Nippon Kagaku Zasshi 76:9025. Japan Industrial Standard (JIS) (1974) K01026. Shull KE, Guthan GR (1967) J Am Water Works Assoc 59:1456

References 289

7. Titkova NF (1963) Nauchn Doklady Vyssn Shk, Biol Nauki 126:1278. Simons LH, Monaghan PH, Taggart MS (1953) Anal Chem 25:9899. Sackett W, Arrhenius G (1962) Geochim Cosmochim Acta 26:955

10. Nishikawa Y, Hiraki K, Morishige K, Tsuchiyama A, Shigematsu T (1968) BunsekiKagaku 17:1092

11. Shigematus T, Nishikawa Y, Hiraki K, Nagano N (1970) Bunseki Kagaku 19:55112. Hydes DJ, Liss PS (1976) Analyst 101:92213. Morishige K, Hiraki K, Nishikawa K, Shigematsu T (1978) Bunseki Kagaku 27:10914. Ishibashi M, Fujinaga T (1952) Nippon Kagaku Zasshi 73:78315. Hsu DY, Pipes WO (1972) Environ Sci Tech 6:64516. Kahn HL (1973) Environ Anal Chem 3:12117. Barnes RA (1975) Chem Geol 15:17718. Lee ML, Burrell DC (1973) Anal Chim Acta 66:24519. Gosink TA (1975) Anal Chem 47:16520. Pavlenko LI, Safronova NS (1975) Zh Anal Khim 30:77521. Fujinaga T, Kusaka R, Koyama M et al. (1973) J Radioanal Chem 13:30122. Buono JA, Karin RW, Fasching JL (1975) Anal Chim Acta 80:32723. Dougan WK, Wilson AL (1974) Analyst 99:41324. Dale T, Henriksen A (1975) Vatten 31:9125. Henriksen A, Bergmann-Paulsen LM (1975) Vatten 31:33926. Korenaga T, Motomizu S, Toei K (1980) Analyst 105:32827. Howard AG, Coxhead AJ, Potter A, Watt AP (1986) Analyst 111:137928. Salgado Ordonez M, Garcia de Torres A, Cano Pavon JM (1985) Talanta 32:88729. Spencer DW, Sachs PL (1969) Atomic Absorption Newsletter 8:6530. Van den Berg CMG, Murphy K, Riley JP (1986) Anal Chim Acta 188:17731. Prochazkova L (1964) Anal Chem 36:86532. Johnston R (1966) Journal of the International Conference on Exploration of the Sea

N:1133. Strickland JDH, Austin KH (1959) Journal of the International Council for Exploration

of the Sea 24:44634. Newell BS (1967) J Mar Biol Assoc 47:27135. Emmet RT (1968) Naval Ship Research and Development Centre, Report 257036. Koroleff F (1970) Interlaboratory Report 3:1937. Solorzano L (1969) Limnol Oceanogr 14:79938. Head PC (1971) Deep-Sea Res 18:53139. Grasshoff K, Johansson H (1972) Journal of the International Council for Exploration

of the Sea 34:51640. Liddicoat MI, Tibbits S, Butler EI (1975) Limnol Oceanogr 20:13141. Slawyk G, Maclsaac JJ (1972) Deep-Sea Res 19:52142. Benesch R, Mangelsdorf P (1972) Helgolander wiss Meeresunters 23:36543. Haywood JE, Huyser DJ (1970) Water Res 4:69544. Matsunaga K, Nishimura M (1971) Japan Analyst 29:99345. Richards FA, Kletsch RA (1964) In: Maysho Y, Kryama T (eds) Recent researches in

the fields of hydrosphere, atmosphere, and nuclear geochemistry. Maruzen, Tokyo, pp65, 81

46. Matsunaga K, Nishimura M (1974) Anal Chim Acta 73:20447. Berg BR, Abdullah MI (1977) Water Res 11:63748. Brzezinski MA (1987) Mar Chem 20:27749. Le Corre P, Treguer P (1978) Journal of the International Council for Exploration of

the Sea 38:147


50. Grasshoff K, Johannsson H (1972) Journal of the International Council for Explorationof the Sea 34:516

51. Truesdale VW (1971) Analyst 96:58452. Selmer JS, Sorenson F (1986) Appl Environ Microbiol 52:57753. Willason SW, Johnson KS (1986) Mar Biol 91:28554. McLean JD, Stenger VA, Reim RE, Long MW, Hiller TA (1978) Anal Chem 50:130955. Gilbert TR, Clay AM (1973) Anal Chem 45:175756. Hampson BL (1977) Water Res 11:30557. Gardner WS, St John PA (1991) Anal Chem 63:53758. Gardner WS, Herche LR, ST John PA, Seitzinger SP (1991) Anal Chem 63:183859. Sturgeon RE, Willie SN, Berman SS (1985) Anal Chem 57:231160. BertineKK,LeeDS(1981)Tracemetals in seawater. In:WongCSet al. (eds)Proceedings

of a NATO advanced research institute on trace metals in seawater, Sicily, Italy. PlenumPress, New York

61. Afvanse YA, Ryabmin A, Azhipa L (1975) Zhur Anal Chem 30:183062. Arsenic in Potable and Seawater by Spectrophotometry, 1978, Tentative Method, (1980)

Department of the Environment/National Water Council Standing Committee of An-alysts, HMSO, London

63. Howard AG, Comber SDW (1992) Microchim Acta 109:2764. Bermejo-Barrera P, Barciela-Alonso MC, Ferrón-Novais M, Bermejo-Barrera A (1995)

J Anal At Spectrom 10: 24765. Burton JD (1981) Trace metals in seawater. In: Wong CS et al. (eds) Proceedings of

a NATO advanced research institute on trace metals in seawater, Sicily, Italy. PlenumPress, New York

66. Ryabinum AI, Romanov AS, Khatakov SL, Kist AA, Khamidove R (1972) Zhur AnalChem 27:94

67. Yusov AN, Ishan ZB (1994) J Radioanal Nucl Chem 179:27768. Creed JT, Magnusson ML, Brockhoff CA, Chamberlain I, Sivaganesan M (1996) J Anal

At Spectrom 11:50569. Klane B, Blum JD (1999) Anal Chem 71:140870. Jaya S, Rao TP, Rao GP (1987) Talanta 34:57471. Hua C, Jagner D, Renman L (1987) Anal Chim Acta 207:26372. Huiliang H, Jagner D, Renum L (1988) Anal Chim Acta 207:3773. Becker NSC, McRae VM, Smith JD (1985) Anal Chim Acta 173:36174. Chow TJ, Goldberg ED (1961) Geochim Cosmochim Acta 20:19275. Bishop JKB (1990) Anal Chem 62:55376. Epstein MS, Zander AT (1979) Anal Chem 51:91577. Roe KK, Froelich PN (1984) Anal Chem 56:272478. Dehairs F, De Bondt M, Baeyens W, Van dr Winkle P, Hoenig M (1987) Anal Chim Acta

196:3379. Okutani T, Tsuruta Y, Sakuragawa (1993) Anal Chem 65:127380. Matahara G (1974) Bunseki Kagaku 23:87881. Yamamuchi N, Nishida T, Nishida H (1989) Bunseki Kagaku 34:4882. Tao H, Imazawa T, Myazaki, Bansho K (1987) Bunseki Kagaku 36:44783. Nishida H (1990) Bunseki Kagaku 39:80584. Fleet B, Liberty KV, Wets TS (1970) Talanta 17:20385. Terashima S (1973) Bunseki Kagaku 22:131786. Matsuzaki K (1975) Bunseki Kagaku 24:44287. Terashima S (1982) Bunseki Kagaku 31:72788. Asani T, Fukuzawa F (1985) Soil Sci Plant Nutr 31:43

References 291

89. Shimomura S, Morita H, Kuboto M (1986) Bunseki Kagaku 25:53990. Nakajima R (1978) Bunseki Kagaku 37:18591. Paschal DC, Bailey GG (1986) Atom Spectrosc 7:192. Schmidt WF, Diotc F (1987) Fresen Z Anal Chem 326:4093. Elles WG, Hodge VR, Darley DA, Jones CL (1988) Atom Spectrosc 9:18194. Shon X-Q, Yian Z, Ni Z-M (1989) Anal Chim Acta 217:27195. Shijo Y, Mitsuhashi M, Shimizu T, Sakunai S (1992) Analyst 117:192996. Soo LD (1982) Anal Chem 54:168297. Gilbert TJ, Hume DN (1973) Anal Chim Acta 65:45198. Florence TM (1974) J Electroanal Chem 49:25599. Eskilsson H, Jaguer D (1982) Anal Chim Acta 138:27

100. Gillain G, Duychaerts G, Distech A (1979) Anal Chim Acta 106:23101. Jagner D, Aren K (1978) Anal Chim Acta 100:375102. Danielsson LG, Jagner D, Josefson M, Westerlund S (1981) Anal Chim Acta 127:147103. Jagner D, Josefson M, Westerlund S (1981) Anal Chim Acta 128:155104. Portman JE, Riley JP (1966) Anal Chim Acta 34:201105. Koroleff F (1947) Acta Chim Scand 1:503106. KoroleffF (ed) (1980)Proceedingsof the12thConferenceof theBalticOceanographers,

Leningrad107. Komorsky-Lovric S (1988) Anal Chim Acta 204:161108. Uppstroem LR (1968) Anal Chim Acta 43:475109. Hulthe P, Uppstroem LR, Oestling G (1970) Anal Chim Acta 51: 31110. Nicolson RA (1971) Anal Chim Acta 56:147111. Oshima M, Motomizu S, Toei K (1974) Anal Chem 56:948112. Marcantoncetos M, Gamba G, Mannier D (1973) Anal Chim Acta 67:220113. Horta TMTC, Curtins AJ (1977) Anal Chim Acta 96:207114. Tsaikov SP (1982) Comptes Rendus De I’Academic Bulgare Des Science 35:61115. Campbell WC, Ottaway JH (1977) Analyst 102:495116. Guevremont R (1980) Anal Chem 52:1574117. Guevremont R, Sturgeon RE, Berman SS (1980) Anal Chim Acta 115:1633118. Gardner J, Yates J (1979) In: International conference on management and control

of heavy metals in the environment. Water Research Centre, Stevenage Laboratory,Stevenage. CEP Consultants, Edinburgh, pp 427–430

119. Danielsson LG, Magnusson B, Westerlund S (1978) Anal Chim Acta 98:47120. Batley GE, Farrah YJ (1978) Anal Chim Acta 99:283121. Sturgeon RE, Berman SS, Desauiniers JAH, Russell DS (1980) Talanta 27:85122. Bruland KW, Franks RP, Knauer GA, Martin JH (1979) Anal Chim Acta 105:233123. Smith RG Jr, Windom HL (1980) Anal Chim Acta 113:39124. Rasmussen L (1981) Anal Chim Acta 125:117125. Sturgeon RE, Berman SS, Desauiniers JAH, Mykytiuk AP, McLaren JW, Russell DS

(1980) Anal Chem 52:1585126. Sperling KR (1978) Anal Chem 292:113127. Bengtsson M, Danielsson LG, Magnusson B (1979) Anal Letters 12:1367128. Sperling KR (1980) Anal Chem 301:294129. Kingston HM, Barnes IL, Brady TJ, Rains TC, Champ MA (1978) Anal Chem 50:2064130. Lund W, Larsen BV (1974) Anal Chim Acta 72:57131. Bately GE (1981) Anal Chim Acta 124:121132. Lundgren G, Lundmark L, Johansson G (1974) Anal Chem 46:1028133. Sperling KR (1977) Fresen Z Anal Chem 287:23134. Sperling KR (1980) Fresen Z Anal Chem 301:294


135. Pruszkowska E, Carnrick GR, Slavin W (1983) Anal Chem 55:182136. Campbell WC, Ottaway JM (1977) Analyst 102:495137. Nakahara T, Chakrabarti CL (1979) Anal Chim Acta 104:99138. Brewer SW (1985) Anal Chem 57:724139. Knowles M (1987) Varian atomic absorption no AA 71 methods for the determination

of cadmium in seawater with Zeeman background correction140. Lum KR, Callaghan M (1986) Anal Chim Acta 187:159141. Zhang Z, Liu J, Lin R, Yang X, He H (1988) Zhongshan Daxue Xuebao Ziran Kexuebao

3:109142. Han H, Le X, Ni Z (1986) Huanjing Huaxue 5:34143. Stolzberg RJ (1977) Anal Chim Acta 92:193144. Yoshimura W, Uzawa Z (1987) Bunseki Kagaku 36:367145. Kounaves SP, Zirino A (1979) Anal Chim Acta 109:327146. Turner DR, Robinson SG, Whitfield M (1984) Anal Chem 56:2387147. Frigieri P, Trucco R, Ciaccolini I, Pampurini G (1980) Analyst 105:651148. Ganzerli L, Maggi L, Ciceri G (1987) J Radioanal Nucl Chem 114:105149. Shen Z, Li P (1986) Fenxi Huaxue 14:55150. Jagner D (1974) Anal Chim Acta 68:83151. Schmid RW, Reilley CN (1957) Anal Chem 29:264152. Ringbom AG, Pensar G, Wanninen E (1958) Anal Chim Acta 19:525153. Sadek FS, Schmid RW, Reilley CN (1959) Talanta 2:38154. Wanninen E (1961) Talanta 8:355155. Date Y, Toel K (1963) Bull Chem Soc Japan 36:518156. Culkin F, Cox RA (1966) Deep-Sea Res 13:789157. Tsunogai S, Nishimura M, Nakaya S (1968) Talanta 15:385158. Schwarzenbach G, Flaschka H (1969) Complexometric titrations, 2nd ed. Methuen,

London159. Horibe Y, Endo K, Tsuboto H (1974) Earth Planet Sci Lett 23:136160. Jagner D (1974) Anal Chim Acta 68:83161. Lebel J, Poisson A (1976) Mar Chem 4:321162. Krumgalz D, Holzer R (1980) Limnol Oceanogr 25:367163. Kanamori S, Ikegami H (1980) J Oceanographic Soc Japan 36:177164. Olson EJ, Chen CTA (1982) Limnol Oceanogr 27:375165. Smith MR, Cockram HP (1981) Atom Spect 2:97166. Dybus J (1969) Clin Chim Acta 23:309167. Blake WE, Bryant MWR, Waters A (1969) Analyst 94:49168. Whitfield M, Leyendekkers JW, Kerr JD (1969) Anal Chim Acta 45:399169. Brenner JB, Eldad H, Erlich S, Dalman N (1984) Anal Chim Acta 166:51170. Nakayama E, Kuwamoto T, Tokoro H, Fujinaka T (1981) Anal Chim Acta 131:247171. Cranston RE, Murray JW (1978) Anal Chim Acta 99:275172. Dejong GJ, Brinkman UAT (1978) Anal Chim Acta 98:243173. Gilbert TR, Clay AM (1973) Anal Chim Acta 67:289174. Crosmun ST, Mueller TR (1975) Anal Chim Acta 75:199175. Batley GE, Matousek JP (1980) 5 Anal Chem 52:1570176. Pankow JF, Junauer GE (1974) Anal Chim Acta 69:97177. Miyazaki A, Barnes RM (1981) Anal Chem 53:364178. Willie SN, Sturgeon RE, Berman SS (1983) Anal Chem 55:981179. Moffett JH (1987) Varian Atomic Absorption No. AA69, Measurement of Varain

Techitron Pty, Ltd., Mulgrave, Victoria, Australia. Chromium in environmental watersby Zeeman corrected graphite tube atomisation

References 293

180. Dubovenko LI, Zaposlozets QA, Pyatnitskii II (1986) Khim Teknol Vody 8:50181. Siu WM, Bednas HE, Berman SS (1983) Anal Chem 55:473182. Heumann KG (1980) Toxicol Environ Chem Rev 3:111183. Colby BN, Rosecrance AE, Colby ME (1981) Anal Chem 53:1907184. Mykytiuk AP, Russell DS, Sturgeon RE (1980) Anal Chem 52:1281185. Sturgeon RE, Berman SS, Williw SN, Desauiniers JAH (1981) Anal Chem 53:2337186. Boussemart M, Van den Berg CMG (1994) Analyst 119:1349187. Chang CA, Patterson HH, Mayer LM, Bause DE (1980) Anal Chem 52:1264188. Cranston RE, Murray JW (1980) Limnol Oceanog 25:1104189. Mullins RL (1984) Anal Chim Acta 165:97190. Sandell EB (1959) Calorimetric determination of traces of metals. Interscience, New

York191. Marchart H (1984) Anal Chim Acta 30:11192. Willems GJ, de Ranter CJ (1974) Anal Chim Acta 68:111193. Myasoedova GV, Savvin SG (1982) Zh Anal Khim 37:499194. Leyden DE, Wegscheider W (1981) Anal Chem 53:1059195. Ahern F, Eckert JM, Payne NC, Williams KL (1985) Anal Chim Acta 175:147196. Posta J, Alimonti A, Petrucci F, Caroli S (1996) Anal Chim Acta 325:185197. Mugo RK, Orlans KJ (1992) Anal Chim Acta 271:1198. Nakayama E, Kuwamoto H, Tokoro H, Tsurubo S, Fujinaga T (1981) Anal Chim Acta

130:289199. Grimaud D, Michard G (1974) Mar Chem 2:229200. Nakayama E, Kuwamoto H, Tokoro H, Tsurubo S, Fujinaga T (1981) Anal Chim Acta

130:401201. Nakayama E, Kuwamoto H, Tokoro H, Tsurubo S, Fujinaga T (1981) Anal Chim Acta

131:247202. Ishibashi M, Shigematsu T (1950) Bulletin of the Institute of Chemical Research Kyoto

University 23:59203. Cheucas L, Riley JP (1966) Anal Chim Acta 35:240204. Fukai R (1967) Nature 213:901205. Fukai R, Vas D (1967) J Oceanographic Society Japan 23:298206. Kuwamoto T, Murai S (1970) Preliminary report of the Hakuho-maru cruise KH-68-4.

Ocean Research Institute, University of Tokyo, Tokyo p 72207. Yamamoto T, Kadowski S, Carpenter JH (1974) Geochem Journal 8:123208. Bowen HJM (1966) Trace elements in biochemistry. Academic Press, London209. Kendrick MJ, May MT, Plishka MJ, Robinson KD (1992) Metals in biological systems.

Ellis Harwood, Chichester pp 71–79210. Knauer GA, Martin JH, Gordon RM (1982) Nature 297:49211. Bruland KW (1983) Trace elements in seawater. In: Riley JP, Chester R (eds) Chemical

oceanography, vol 8. Academic Press, London212. Fukai R, Murray CN (1973) Environmental behaviour of radiocobalt and radiosilver re-

leased from nuclear power stations into aquatic systems. In: Environmental behaviourof radionuclides released in the nuclear industry. IAEA, Vienna pp 217–242

213. Carvalho FP (1986) Health Physics 53:73214. Nolan CV, Fowler SW, Teyssie JL (1992) Mar Ecol-Prog Ser 30:105215. Turner DR, Whitfield M, Dickson A (1981) Geochim Cosmochim Acta 45:855216. Zhang H, van den Berg CMG, Wollast R (1990) Mar Chem 28:285217. Donat JR, Bruland KW (1988) Anal Chem 60:240218. Huyng Ngoc L, Whitehead NE (1986) Oceanol Acta 9:433219. Jickells TD, Burton JD (1988) Mar Chem 23:131


220. Zhang H, Wollast R (1990) Distributions of dissolved cobalt and nickel in the Rhoneand the Gulf of Lyons. In: Martin JM, Barth H (eds) Water pollution research report20, ERO 2000, 2nd Workshop on the northwest mediterranean sea. Blanes, Spain pp397–414

221. Murray JW (1975) Geochim Cosmochim Acta 39:635222. Balistrieri LS, Murray JW (1975) Geochim Cosmochim Acta 50:2235223. Kentner E, Zeitlin H (1970) Anal Chim Acta 49:587224. Motomizu S (1973) Anal Chim Acta 64:217225. Forster W, Zeitlin H (1966) Anal Chim Acta 34:211226. Riley J, Topping G (1969) Anal Chim Acta 44:234227. Armitage B, Zeitlin H (1971) Anal Chim Acta 53:47228. Going J, Wesenberg G, Andrejat G (1976) Anal Chim Acta 81:349229. Korkisch J, Sorio A (1975) Anal Chim Acta 79:207230. Gurtler O (1977) Fresen Z Anal Chem 284:206231. Harvey BR, Dutton JWR (1973) Anal Chim Acta 67:377232. Kouimtzis TA, Apostolopoulou C, Staphilakis I (1980) Anal Chim Acta 113:185233. Vasilikiotis G, Kouimtzis Th, Apostolopoulou C, Voulgaropoulos (1974) Anal Chim

Acta 70:319234. Isshiki K, Nakayama E (1987) Anal Chem 59:291235. Spencer CP (1957) J Gen Microbiol 16:282236. Johnston R (1962) J Mar Biolog Assoc 43:427237. Johnston R (1964) J Mar Biolog Assoc 44:87238. Barber RT, Ryther JH (1969) J Exploratory Marine Biol 3:191239. Danielsson LG (1980) Mar Chem 8:199240. Malahoff A, Kolotyrkina IY, Shipgun LK (1996) Analyst 121:1037241. Yuzefovsky A, Leonardo RF, Wang M, Michel RG (1994) J Anal Atom Spectrosc 9:1195242. Boyle EA, Handy B, van Green A (1987) Anal Chem 59:1499243. Sakamoto N private communication244. Hirata S, Hashimoto Y, Aihara M, Mallika GV (1996) Fresen J Anal Chem 355:676245. Vega M, Van den Berg CMG (1997) Anal Chem 69:874246. Sylva RN (1976) Water Res 10:789247. Sunda WG, Hanson AK (1987) Limnol Oceanog 32:537248. Sunda WG, Lewis JM (1978) Limnol Oceanog 23:870249. Sunda WG, Gillespie PA (1979) J Mar Res 27:761250. Lewin J, Chen CH (1971) Limnol Oceanog 16:670251. Lewis AG, Whitfield RH, Ramnarine A (1972) Mar Biol 17:215252. Davey EW, Morgan MJ, Erickson SJ (1973) Limnol Oceanog 18:993253. Prakash A, Rashid MA, Jensen A, Subba Rao DV (1973) Limnol Oceanog 18:516254. Gnassia-Barelli M, Romero M, Laumond F, Pesando D (1978) Mar Biol 47:15255. Hongve D, Skogheim OK, Hindar A, Abrahamsen H (1980) Bull Environ Contam

Toxicol 25:594256. Fisher NS, Frood D (1980) Mar Biol 59:85257. Sunda W, Guillard RRl (1976) J Mar Res 34:511258. Anderson DM, Morel FM (1978) Limnol Oceanog 23:283259. Murphy LS, Guillard RRL, Gavis J (1980) Evolution of resistant phytoplankton strains

through exposure to marine pollutants. In: Mayer GF (ed) Ecological stress and thenew york blight: science and management. Estuarine Research Foundation, Columbia,SC, USA, pp 401–412

260. Jackson GA, Morgan JJ (1978) Limnol Oceanog 23:268261. Chester R, Stoner JH (1974) Mar Chem 2:17

References 295

262. Alberts JJ, Leyden DE, Patterson TA (1976) Mar Chem 4:51263. Boyle FA, Sclater FR, Edmond JM (1977) Earth Planet Sci Lett 37:38264. Winlom HL, Smith RG (1979) Mar Chem 7:157265. Zirino A, Yamaoto S (1972) Limnol Oceanog 17:661266. StummW, Morgan JJ (1981) Aquatic chemistry: an introduction emphasizing chemical

equilibria in natural waters, 2nd edn. Wiley, New York267. Sibley TH, Morgan JJ (1975) Equilibrium speciation of trace metals in freshwater–

seawater mixtures. In: Hutchinson HC (ed) Proceedings of international conferenceon heavy metals in the environment, University of Toronto, Toronto, Ontario pp 310–338

268. Pagenkopf GK, Russo RC, Thurston RV (1974) Journal of the Fisheries Research Boardof Canada 31:462

269. Shuman MS, Woodward GP (1973) Anal Chem 45:2032270. Chau R, Lum-Shue-Chan K (1974) Water Res 8:383271. Campbell PGC, Blisson M, Gagne R, Tessior A (1977) Anal Chem 49:2358272. Stiff MJ (1971) Water Res 5:585273. Batley GE, Florence TM (1976) Anal Lett 9:379274. Plasvic M, Krznacic D, Branica M (1982) Mar Chem 11:17275. Van den Berg CMG (1982) Mar Chem 11:323276. Florence TM, Batley GE (1976) Talanta 23:179277. Figura P, McDuffie B (1979) Anal Chem 51:120278. Ruzic I (1982) Anal Chim Acta 140:99279. Waite TD, Morel FMM (1983) Anal Chem 55:1268280. Cabon JY, Le Bihan A (1987) Anal Chim Acta 198:87281. Cabon JY (1987) Anal Chim Acta 198:103282. Nishoika H, Maeda Y, Azumis T (1986) Nippon Kaisui Gakkashi 40:100283. Wang M, Yuzetovsky AJ, Michel RG (1993) Microchem J 48:326284. Tasinski R, Trachtenberg I, Andrychuk D (1974) Anal Chem 46:364285. De Marco R (1994) Anal Chem 66:3202286. Abraham J, Winpe M, Ryan DE (1969) Anal Chim Acta 48:431287. Muzzarelli RAA, Roccheti R (1974) Anal Chim Acta 69:35288. Gukens J (1987) Ion Selective Electrode Review 9:127289. Westall JC, Morel FMM, Hume DN (1979) Anal Chem 51:2106290. Lewenstam A, Hulanicki A, Ghali E (1990) In: Ivanaska I et al. (eds) Contemporary

electron chemistry. Plenum Press, New York291. Milner GWC, Wilson JD, Barnett GA (1961) Journal of Electro Anal Chem 2:25292. Van den Berg CMG, Huang ZQ (1984) Anal Chim Acta 164:209293. Scarano G, Morelli F, Seritti A, Zirino A (1990) Anal Chem 62:943294. Quentel F, Elleouet C, Madec C (1994) Electroanalysis 6:683295. Wang J, Foster N, Armalis S, Larsen D, Zirino A, Olsen K (1995) Anal Chim Acta

310:223296. Garcia-Monco Carra R, Sanchez-Misiego A, Zirino A (1995) Anal Chem 67:4484297. Fairless C, Bard AJU (1973) Anal Chem 45:2289298. Abe T, Itoh K, Murozumi M (1977) Bunseki Kagaku 15:65299. Harvey BR (1978) Anal Chem 50:1866300. Virmani YP, Zeller EJ (1974) Anal Chem 46:324301. Zharikov VF, Senyavin MK (1970) Trudy Gos Okean Inst (101), Ref Zhur Khim 19GD

(7), Abstract No 7G189302. Marvin KT, Proctor RR, Neal RA (1970) Limnol Oceanog 15:320303. Wood AM, Evans DW, Alberts JJ (1983) Mar Chem 13:305


304. Stolzberg RJ, Rosin D (1977) Anal Chem 49:226305. Giesy JP (1980) Cadmium interactions with naturally occuring ligands. In: Nriagu JC

(ed) Biogeochemistry of cadmium, pt. 1. Wiley, New York, pp 237–256306. Zorkin NG, Grill EV, Lewis AG (1986) Anal Chim Acta 183:163307. Zirino A, Healy ML (1970) Limnol Oceanog 15:956308. Bradford WL (1978) Limnol Oceanog 1:757309. Bilinski H, Huston R, Stumm W (1976) Anal Chim Acta 84:157310. O’Shea TA, Mancy KH (1976) Anal Chem 48:1603311. Duinker JC, Kramer CJM (1977) Mar Chem 5:207312. Piro A, Bernhard M, Branica M, Verzi M (1972) Radioactive contamination of the ma-

rine environment. Symposium proceedings, Seattle, WA, USA. IAEA Edition, Vienna,pp 287–304

313. Ernst R, Allen HE, Mancy KH (1975) Water Res 9:969314. Florence TM, Batley GB (1977) Journal of Electro Anal Chem 75:791315. Stumm W, Bilinski H (1973) Advances in water pollution research. Sixth international

conference. Jerusalem, 1972. Pergamon, Oxford, p 39316. Bubic S, Branica M (1973) Thalassia Jugoslavia 9:47317. Zirino A, Kounaves SP (1977) Anal Chem 49:56318. Zirino A, Kounaves SP (1979) Anal Chem 51:592319. Zirino A, Kounaves SP (1977) Anal Chim Acta 113:79320. Sheffrin N, Williams EE (1982) Anal Proc 19:483321. Jagner D, Granelli A (1976) Anal Chim Acta 83:19322. Jagner D (1978) Anal Chem 50:1924323. Jagner D, Aren K (1979) Anal Chim Acta 107:29324. Brooks RR, Presley BJ, Kaplan IR (1967) Anal Chim Acta 38:321325. Brooks RR, Presley BJ, Kaplan IR (1967) Talanta 14:809326. Shuman MS, Michael LC (1975) In: Proceedings of the international conference on

heavy metals in the environment, vol 1. Wiley, New York, p 227327. Shuman MS, Michael LC (1978) Environ Sci Tech 12:1069328. Odier M, Plichon V (1971) Anal Chim Acta 55:209329. Weiss HV, Kenis PR, Korkisch J, Steffan I (1979) Anal Chim Acta 104:337330. Van den Berg CMG (1984) Anal Chim Acta 164:195331. Turner DB, Robinson SG, Whitfield M (1984) Anal Chem 56:2387332. Sunda WG, Hanson AK (1987) Limnol Oceanog 325:37333. Andreae MO, Froelich P (1981) Anal Chem 53:287334. Falkner KK, Edmond JM (1990) Anal Chem 62:1477335. Pilipenko AT, Pavlova VK (1972) Zh Anal Khim 27:1253336. Matthews AD, Riley JP (1970) Anal Chim Acta 51:287337. Martin JH, Fitzwater SE (1988) Nature 331:342338. Martin JH, Gordon RM, Fitzwater SE, Broenkow WW (1989) Deep-Sea Res 36:649339. Martin JH, Gordon RM, Fitzwater SE (1990) Nature 345:156340. Martin JH, Gordon RM, Fitzwater SE (1990) Global Biogeochemical Cycles 4:5341. Martin JH (1991) Nature 353:123342. Betzer PR, Carder KL, Duce RA, Merrill JT, Tindale NW, Uematsu M, Costello DK,

Young RW, Feely RA, Breland JA, Bernstein RE, Greco AM (1988) Nature 336:568343. Young RW, Karter K, Betzer PR (1991) Global Biogeochemical Cycles 5:119344. Broecker WS (1990) Global Biogeochemical Cycles 4:13345. Dugdale RC, Wilkinson FP (1990) Global Biogeochemical Cycles 4:13346. Zhuang G, Yi Z, Duce RA, Brown PR (1992) Nature (London) 355:537347. Rigin VL, Blotchin AJ (1977) Zh Anal Khim 32:312

References 297

348. Pilipenko AT, Barovskii VA, Kalinichenko JE (1978) Zh Anal Khim 33:1880349. Obata H, Karatani H, Nakayama E (1993) Anal Chem 65:1524350. Shriadah MMA, Ohzeki K (1986) Analyst 111:555351. Blain S, Treguer P (1995) Anal Chim Acta 308:425352. Eirod VA, Johnson KS, Coale KH (1991) Anal Chem 63:893353. O’Sullivan DW, Hansson AK, Lester DR (1995) Mar Chem 49:65354. Hiiro K, Tanaka T, Sawada T (1972) Japan Analyst 21:635355. Moore RM (1981) Trace metals in seawater. In: Wong CS et al. (eds) Proceedings of


356. Brendel PJ, Luther III GW (1995) Environ Sci Tech 29:751357. Sharma GM, Du Bois HR (1978) Anal Chem 50:516358. Bolshov MA, Zybin AV, Koloshnikov VG, Vasnetsov MV (1981) Spectrochim Acta

36B:345359. Cheam V, Lechner J, Sekerka I, Desrosiers R (1994) J Anal Atom Spectrosc 9:315360. Ke C-B, Lin K-C (1999) Anal Chem 71:1561361. Frech W, Cedergreu A (1977) Anal Chim Acta 88:57362. Halliday MC, Houghton C, Ottaway JM (1986) Anal Chim Acta 119:67363. Schaule BK, Patterson CC (1981) Earth Planet Sci Lett 54:97364. Hirao Y, Fukomoto K, Sugisaki H, Kimura K (1979) Anal Chem 51:651365. Torsi G, Desimoni E, Palmisano F, Sabbatini L (1981) Anal Chim Acta 124:143366. McLaren JW, Wheeler RC (1977) Analyst 102:542367. Adams MJ, Kirkbright GF, Rienvatana P (1977) Atomic Absorption Newsletter 14:10368. Culver B, Surles T (1976) Anal Chem 47:920369. Prithard MW, Reeves RD (1976) Anal Chim Acta 82:103370. Yasuda S, Kakihama H (1976) Anal Chim Acta 84:291371. Tessari G, Torsi G (1975) Anal Chem 47:842372. Sturgeon RE, Chakrabarti CL, Langford CN (1976) Anal Chem 48:1792373. Fernandez FJ, Manning DC (1971) Atomic Absorption Newsletter 10:3374. Ediger RD, Peterson G, Kerber JD (1974) Atomic Absorption Newsletter 13:61375. Ediger RD (1975) Atomic Absorption Newsletter 14:127376. Regan JGT, Warren J (1976) Analyst 101:220377. Frech W, Cedergren A (1977) Anal Chim Acta 88:57378. Manning DC, Slavin W (1978) Anal Chem 50:1234379. Lundgren G, Lundmark L, Johansson G (1974) Anal Chem 46:1028380. Scobbie R (1973) Technical topics. Varian Techtron, Springvale, Victoria, Australia381. Donnelly TH, Ferguson J, Eccleston AJ (1975) Appl Spectros 29:2382. Shigematsu T, Matsui M, Fujino O, Kinoshita K (1975) Anal Chim Acta 76:329383. Branderberger H, Bader H (1967) Atomic Absorption Newsletter 6:101384. Lund W, Larsen BV (1974) Anal Chim Acta 70:299385. Lund W, Larsen BV (1974) Anal Chim Acta 72:57386. Lund W, Larsen BV, Gundersen N (1976) Anal Chim Acta 81:319387. Lund W, Thomassen Y, Doyle P (1977) Anal Chim Acta 93:53388. Fairless L, Bard AJ (1972) Anal Lett 5:433389. Fairless L, Bard AJ (1973) Anal Chem 45:2289390. Batley GE, Matousek JP (1977) Anal Chem 49:2031391. Jensen PO, Dolezal J, Langmyhr FJ (1975) Anal Chim Acta 72:245392. Newton MP, Chauvin JV, Davis DG (1973) Anal Lett 6:89393. Newton MP, Davis DG (1975) Anal Chem 47:2003394. Dawson JB, Ellis DJ, Hartley TF, Evans MEA, Metcalf KW (1974) Analyst 99:602


395. Torsi G (1977) Anal Chim (Rome) 67:557396. Halliday MC, Houghton C, Ottaway JM (1980) Anal Chim Acta 119:67397. Ohta K, Suzuki M (1979) Fresen Z Anal Chem 298:140398. Sturgeon RE, Willie SN, Berman SS (1989) Anal Chem 61:1867399. Clem RG, MDI Application, University of California, Berkeley. Laboratory Notes 8:1400. Petrie LM, Baier RW (1978) Anal Chem 50:351401. Acebal SA, De Luca Rebello A (1983) Anal Chim Acta 148:71402. Quentel F, Madec C, Courtot-Coupez J (1988) Water Res 20:325403. Svensmark B (1987) Anal Chim Acta 197:239404. Wu Q, Batley GE (1995) Anal Chim Acta 309:95405. Garcia-Monco Carra R, Sanchez-Misiego A, Zirino A (1995) Anal Chem 67:4484406. Flegal AR, Stukas VJ (1987) Mar Chem 22:163407. Byrne RH (1981) Nature 290:487408. Scaule DM, Patterson CC (1981) Trace metals in seawater. In: Wong CS et al. (eds)

Proceedings of a NATO advanced research institute on trace metals in seawater, Sicily,Italy. Plenum Press, New York

409. Benzwi N (1972) Israel J Chem 10:967410. Rona M, Schmuckler G (1973) Talanta 20:237411. Wiernik M, Amiel S (1970) J Radioanal Chem 5:123412. Das DK, Mazumdar M, Shome SC (1972) Anal Chim Acta 60:439413. Smith MR, Cockran HB (1981) Atom Spectrosc 2:97414. Pybus T (1969) Anal Chim Acta 23:309415. Anfält T, Granéli A Anal Chim Acta 86:13416. Stumm W, Morgan JJ (1970) Aquatic chemistry. Wiley Interscience, New York417. Murray JW, Brewer PG (1977) In: Glasby GP (ed) Marine manganese deposits. Elsevier,

Amsterdam, pp 291–325418. Morgan JJ (1967) Chemical equilibria and kinetic properties of manganese in natural

waters. In: Faust SD, Hunter JV (eds) Principles and applications of water chemistry,Wiley, New York, pp 561–624

419. Kessick MA, Morgan JJ (1975) Environ Sci Tech 9:157420. Stumm W, Giovanoli R (1976) Chimica 30:423421. Bryan GW, Hummerstone IG (1973) J Mar Biol Assoc of the UK 53:705422. Holliday LM, Liss PS (1976) Estuar Coast Mar Sci 4:349423. Evans DW, Cutshall NH, Cross FA, Wolfe DA (1977) Estuar Coast Mar Sci 5:71424. Sholkovitz ER (1978) Earth Planet Sci Lett 41:77425. Davison W (1976) J Electroanal Chem 72:229426. Davison W (1977) Limnol Oceanogr 22:746427. Olafsson J (1986) Sci Total Environ 49:101428. Brewer PG, Spencer DW (1971) Limnol Oceanogr 16:107429. Biddle VI, Wehry EL (1971) Anal Chem 50:867430. Sturgeon RE, German SS, Desaulniers A, Russell DS (1979) Anal Chem 51:2364431. Segar DA, Gonzalez JG (1972) Anal Chim Acta 58:7432. Klinkhammer GP (1980) Anal Chem 52:117433. McArthur JM (1977) Anal Chim Acta 93:77434. Hydes DJ (1980) Anal Chem 52:959435. Segar DA (1975) Adv Chem Ser 147:56436. Segar DA, Cantillo AY (1980) Anal Chem 52:1766437. Montgomery JR, Peterson GN (1980) Anal Chim Acta 117:397438. Manning DC, Slavin W (1978) Anal Chem 50:1234439. L’vov BV (1978) Spectrochim Acta B 33:153

References 299

440. Segar DA, Cantillo AY (1976) Spectrochimica Symposium, American Society of Lim-nology and Oceanography 2:171

441. Slavin W, Manning DC (1979) Anal Chem 51:261442. Manning DC, Slavin W (1980) Anal Chim Acta 118:301443. Slavin W, Manning DC (1980) Spectrochim Acta B 35:701444. Landing WM, Bruland KW (1980) Earth Planet Sci Lett 49:45445. Moore RM, Burton JD, Williams PJ le B, Young ML (1979) Cosmochim Acta 43:919446. Subramanian KS, Meranger JC (1979) Int J Environ Anal Chem 7:25447. Sugimae A (1980) Anal Chim Acta 121:331448. Statham PJ (1985) Anal Chim Acta 169:149449. International Council for Exploration of the Sea (1982) Marine environment quality

committee report CM1983/E24. Charlottenlund, Denmark450. Bender ML, Klinkhammer GP, Spencer DW (1977) Deep-Sea Res 24:799451. Klinkhammer GP, Bender ML (1980) Earth Planet Sci Lett 46:361452. Carnrick GR, Slavin W, Manning DC (1981) Anal Chem 53:1866453. Lan CR, Alfassi ZB (1994) Analyst 119:1033454. Knox S, Turner DR (1980) Estuar Coast Mar Sci 10:317455. Department of the Environment and the National Water Council, Standing Committee

of Analysts (1978) Manganese in raw and potable waters by spectrophotometry (usingformaldoxime), 1977. Methods for the examination of waters and associated materials.HMSO, London

456. Wiggins PT, Duffy D, El Kady AA (1972) Anal Chim Acta 61:421457. Hadeishi T, McLauglin RD (1971) Science 174:404458. Hatch WR, Ott WL (1968) Anal Chem 40:2085459. Hinckle ME, Learned RE (1969) US Geological Survey, Professional papers 650-D:251460. Fishman MJ (1970) Anal Chem 42:1462461. Lindsteat G (1970) Analyst 95:264462. Chau YK, Saitoh H (1970) Environ Sci Tech 4:839463. Omang SH (1971) Anal Chim Acta 53:415464. Omang SH, Paus PE (1971) Anal Chim Acta 56:393465. Hwang JH, Ullucci PA, Malenfant AL (1971) Canadian Spectroscopy 16:1466. Aston SR, Riley JP (1972) Anal Chim Acta 59:349467. Muscat VI, Vickers TJ, Andren A (1972) Anal Chem 44:218468. Topping G, Pirie JM (1972) Anal Chim Acta 62:200469. Carr RA, Hoover JB, Wilkniss PE (1972) Deep-Sea Res 19:747470. Omang SH (1973) Anal Chim Acta 63:247471. Gardner D, Riley JP (1973) Nature 241:526472. Olafsson J (1974) Anal Chim Acta 68:207473. Muscat VI, Vickers TJ (1970) Anal Chim Acta 18:2275474. Leatherland TM, Burton JD, McCartney MJ, Culkin F (1971) Nature 232:112475. Chester R, Gardner D, Riley JP, Stoner J (1973) Mar Pollut Bull 28:4476. Fitzgerald WF, Hunt CD (1973) Abstract of the bulletin of the French oceanographic

union, A-5477. Fitzgerald WF, Lyons WB, Hunt CD (1974) Anal Chem 46:1882478. Olafsson J (1982) Mar Chem 11:129479. Filippelli M (1984) Analyst 109:515480. Blake S (1985) Method for the determination of low levels of mercury in fresh and

saline waters. UE Technical Repor TR229. Water Research Centre, Medmenham481. Gill GA, Fitzgerald WF (1987) Mar Chem 20:277482. Bloxam MG, Gachanja A, Hill SJ, Worsfold PJ (1996) J Anal Atom Spectrosc 11:145


483. Turyan I, Mandler D (1993) Nature 362:703484. Debrak E, Denoyer ER, Tyson JF (1993 ) J Anal Atom Spectrosc 11:127485. Wrembel HZ (1983) Talanta 30:481486. Wrembel HZ, Pajak WA (1986) Chem Anal (Warsaw) 31:409487. Voyce D, Zeitlin H (1974) Anal Chim Acta 69:27488. Agemian H, Da Silvia JA (1979) Anal Chim Acta 104:285489. Davies IM, Pirie JM (1978) Mar Pollut Bull 9:128490. Topping G, Pirie JM, Graham WC, Shepher RJ (1975) An examination of the heavy

metal levels in muscle, kidney, and liver of saithe in relation to year class, area ofsampling, and season. ICE5CM E:37

491. Watling RJ (1975) Anal Chim Acta 75:281492. Nurnberg HW, Valenta P, Sipos L, Branica M (1980) Anal Chim Acta 115:25493. Hiyai Y, Murao Y (1972) Analyst 21:608494. Weiss HV, Crozier TE (1972) Anal Chim Acta 58:231495. Jaya S, Rao TP, Rao GP (1985) Analyst 110:1361496. Pelizzetti E (1997) Water Science Technol Lab 28:1497. Kawabuchi K, Kuroda R (1969) Anal Chim Acta 46:23498. Kuroda R, Tarui T (1974) Fresen Z Anal Chem 269:22499. Korkisch J, Godl L, Gross H (1975) Talanta 22:669500. Kiriyama T, Kuroda R (1984) Talanta 31:472501. Nutaksuka I, Nishimura A, Ohzeki K (1995) Anal Chim Acta 304:243502. Van den Sloot HA, Wals GD, Das HA (1977) Prog Water Technol 8:193503. Muzzarelli RAA, Rochetti R (1973) Anal Chim Acta 64:371504. Emerick RJ (1987) Atom Spectrosc 8:69505. Kuroda R, Matsumoto N, Ogmura K (1988) Fresen Z Anal Chem 330:111506. Specht AA, Beauchemin D (1988) Anal Chem 70:1036507. Hua C, Jagner D, Renman L (1987) Anal Chim Acta 192:103508. Willie SN, Berman SS, Page JA, Van Loon (1987) Canadian J Chem 65:957509. Hidalgo JL, Gomez MA, Caballero M, Cela R, Perez-Bustamonte JA (1988) Talanta

35:301510. Van den Berg CMG (1985) Anal Chem 57:1532511. Kimura A, Yoneda A, Maeda Y, Azumi T (1985) Nippon Kaisui Gakkaishi 40:141512. Hanson G, Szalo A, Chasteen ND (1977) Anal Chem 49:461513. Mok WM, Wai CW (1984) Anal Chem 56:27514. Kulathilake AI, Chatt A (1980) Anal Chem 52:828515. Monien H, Bovenkerk R, Kringe KP, Rath D (1980) Fresen Z Anal Chem 300:363516. Shriadah MMA, Katoaka M, Ohzeki K (1985) Analyst 110:125517. Kim YS, Zeitlin H (1971) Separ Sci 6:505518. Boyle EA, Edmond JA (1977) Anal Chim Acta 91:189519. Bruland KW (1980) Sci Lett 47:176520. Boyle EA, Huested, SS, Jones SP (1981) Journal of Geographical Research 86:8048521. Kentner E, Armitage D, Zeitlin H (1970) Anal Chim Acta 45:343522. Yatsimirskii KB, Ewel’yakov EM, Pablova VK, Savichenko YaS (1973) Okeanologiya

1111, 10. Ref Zhur Khim 19GD Abstract No HG203523. Rampon H, Cavelier R (1972) Anal Chim Acta 60:226524. Lee DS (1982) Anal Chem 60:226525. Nishocka H, Assadamongkol S, Maeda Y, Azumi T (1987) Nippon Kaisui Gakkaishi

40:286526. Van den Berg CMG, Nimmo M (1987) Sci Total Environ 60:185527. Lo J, Cheng BJ, Tseng Cl, Lee JD (1993) Anal Chim Acta 281:429

References 301

528. Koide M, Goldberg ED, Walker R (1996) Deep-Sea Res Part II 43:53529. Koide M, Goldberg ED, Walker R (1996) Deep-Sea Res Part II 43:55530. Wang Z, Li Y, Li Y, Guo Y (1994) Anal Lett 27:957531. Van den Berg CMG, Jacinto GS (1988) Anal Chim Acta 211:129532. Torbjoern A, Jaquer D (1973) Anal Chim Acta 66:152533. Torbjoern A, Jaquer D (1972) Anal Abst 23:994534. Marquis G, Lebel J (1981) Anal Lett 14:913535. Marezak M, Ziaga E (1973) Chemica Acta 18:99536. Ward GK (1979) Report No 9966. National Technical Information Service, Springfield,

VA, USA537. Koide H, Hodge V, Yang JS, Goldberg ED (1987) Anal Chem 59:1802538. Matthews AD, Riley JP (1970) Anal Chim Acta 51:455539. Schoenfeld I, Held S (1969) Israel J Chem 7:831540. Smith RC, Pillai KC, Chow TJ, Folson TR (1965) Limnol Oceanogr 10:226541. Lebedev VA, Alvarez UV, Krasnyanskii AV, Shurnupova TI, Golubtsov IV (1987) Vestn

Mosk Univ Ser Khim 28:1744542. Bryer P, Gilbert BP (1987) Anal Chim Acta 201:33543. Riley JP, Skirrow G (1975) Chemical oceanography, 2nd edn, vol 1. Academic Press,

New York, p 418544. Selenium (1976) Committee on medical and biologic effects of environmental pollu-

tants. National Academy of Sciences, Washington, DC545. Luckey TD, Venegopal B (1976) Chem Eng News 54:2546. Tzeng J-H, Zeitlin H (1978) Anal Chim Acta 101:71547. Neve J, Hanocq M, Molle L (1980) Int J Environ Anal Chem 8:177548. Cutter GA (1978) Anal Chim Acta 98:59549. Goldschmidt VM, Strock LW (1935) Nachr Akad Wiss Goettingen Math-Physik KI, IV,

NF 1:123550. Wattenherg H (1938) Zeitung Anorganic Chemie 236:339551. Ishibashi M, Shigematsu T, Nakagawa Y (1953) Recent oceanographic work in Japan.

Special Number 1:44552. Chau YK, Riley JP (1965) Anal Chim Acta 33:36553. Schutz DF, Turekian KK (1963) Abstracts of papers XIII, Paper 23. General Assembly

IUGG, Berkeley554. Schutz DF (1964) PhD Thesis. Department of Geology, Yale University555. Shimoishi Y (1973) Anal Chim Acta 64:465556. Measures CI, Burton JD (1980) Anal Chim Acta 120:177557. Bermejo-Barrera P, Moreda-Pineiro J, Moreda-Pineiro A, Bermejo-Barrera A (1996)

Talanta 43:35558. Miller LA, Bruland KW (1995) Environ Sci Tech 29:2616559. Kawabuchi K, Riley JP (1973) Anal Chim Acta 65:271560. Kulev II, Stanev DS, Bakalov VD (1974) Khimiya Indust 46:65561. Kulev II, Bakalov VD (1973) CR Bulgarian Acad Science 26:787562. Carr RA (1970) Limnol Oceanogr 15:318563. Petit L (1985) J Internationale d’Oceanographie Medical 19:79564. Andreae MV (1984) Anal Chem 56:2064565. Moran SB, Hoff JA, Buesseler KO, Edwards RL (1995) Geophys Res Lett 22:2589566. Guo L, Santschi PH, Daskaran M, Zindler A (1995) Earth Planet Sci Lett 133:117567. Kodama Y, Tsubota H (1971) Japan Analyst 20:1554568. Dogan S, Haerdi W (1980) Int J Environ Anal Chem 8:249569. Bergerioux C, Haerdi W (1980) Analusis 8:169


570. Brinckmann FE (1981) Trace metals in seawater. In: Wong CS et al. (eds) Proceedingsof a NATO advanced research institute on trace metals in seawater, Sicily, Italy. PlenumPress, New York, p 488

571. Aue WA, Flinn CS (1977) J Chromatogr 142:145572. Jackson JA, Blair WR, Brinckmann FE, Iverson WP (1982) Environ Sci Tech 16:110573. Jewett KL, Brinckmann FE (1981) J Chromatogr Sci 19:583574. Brinckmann FE, Blair WR, Jewett KL, Iverson WP (1977) J Chromatogr Sci 15:493575. Brinckmann FE (1981) J Organomet Chem 12:343576. Jewett KI, Brinckman FE, Bellama JM (1978) Influence of environmental parameters

on transmethylation between aquated metal ions. In: Brinckmann FB, Bellama JM(eds) Organometals and organometalloids: occurrence and fate in the environment.American Chemical Society, Washington, DC, pp 158–187

577. Michael LC, Erickson MD, Parks SP, Pellizari BD (1980) Anal Chem 52:1836578. Braman RS, Tompkins MA (1979) Anal Chem 51:12579. Schaeffer GW, Emilius M (1954) J Am Chem Soc 76:1203580. Hodge VF, Seidel SL, Goldberg FD (1979) Anal Chem 51:1256581. Evans WH, Jackson FJ, Dellar D (1979) Analyst 104:16582. Hallas LE (1978) PhD Thesis. University of Maryland583. Paris GE, Blair WE, Brinckmann FE (1977) Anal Chem 49:378584. Florence TM, Farrar YJ (1974) Journal of Electro Anal Chem 51:191585. Florence TM (1971) Anal Abstr 20:4585586. Florence TM (1972) Anal Abstr 25:5011587. Gandon R, Guegueniat P (1992) Radiochim Acta 57:159588. Yang CY, Shih JS, Yeh YC (1981) Analyst 106:385589. Huh CA, Bacon MP (1985) Anal Chem 57:2065590. Agrawal YK, Upadhyaya DB, Chudasma SP (1993) J Radioanal Nucl Chem 170:79591. Djogic R, Branica M (1995) Analyst 120:1989592. Economou A, Fielden PR, Packham A (1994) Analyst 119:279593. Van den Berg CMG, Nimmo M (1987) Anal Chem 59:924594. Spencer R (1968) Talanta 15:1307595. Hua C, Jagner D, Renman L (1987) Anal Chim Acta 197:265596. Nishimura M, Matsunaga K, Kudo T, Obara F (1973) Anal Chim Acta 65:466597. Kiriyama T, Kuroda R (1972) Anal Chim Acta 62:464598. Monien H, Stangel R (1982) Fresen Z Anal Chem 311:209599. Hastings DW, Emerson SR, Nelson BK (1996) Anal Chem 68:371600. Van den Berg CMG, Huang O (1984) Anal Chem 56:2383601. Vega M, Van den Berg CMG (1994) Anal Chim Acta 293:19602. Weiss HV, Guttman HA, Korkisch J, Stefan I (1977) Talanta 24:509603. Korkisch J, Gross H (1973) Talanta 20:1153604. BrulandKW,FranksRP(1983) In:WongCSetal. (eds)ProceedingsofaNATOadvanced

research institute on trace metals in seawater, Sicily, Italy. Plenum Press, New York, pp256–298

605. Bruland KW, Knauer GA, Martin JH (1978) Nature 271:741606. Danielsson LG, Westerlund S (1983) In: Wong CS et al. (eds) Proceedings of a NATO

advanced research institute on trace metals in seawater, Sicily, Italy. Plenum Press,New York, pp 85–95

607. Magnesson B, Westerlund S (1985) In: Wong CS et al. (eds) Proceedings of a NATOadvanced research institute on trace metals in seawater, Sicily, Italy. Plenum Press,New York, pp 467–473

608. Donat JR, Bruland KW (1990) Mar Chem 28:301

References 303

609. Nowicki JL, Johnson KS, Coale KH, Eirod VA, Lieberman SH (1994) Anal Chem 66:2732610. Burrell DC, Wood GG (1969) Anal Chim Acta 48:45611. Guevremont R (1981) Anal Chem 53:911612. Mykytink AP, Russell DS, Sturgeon RE (1980) Anal Chem 52:123613. Andreae MO, Asmode JF, Foster P, Van’t Dack L (1981) Anal Chem 53:1706614. Andreae MO (1977) Anal Chem 49:820615. Foreback CC (1973) Some studies on the detection and determination of mercury,

arsenic, and antimony in gas discharges. Thesis. University of South Florida616. Huang SD, Shih KY (1995) Spectrochim Acta B 50:837617. Akatsuka K, Katoh T, Nobuyama N, Okanaka T, Okumura M, Hoshi S (1996) Anal Sci

12:109618. Hirata S, Honda K (1987) Bunseki Kagaku 36:213619. Van den Berg CMG (1985) Mar Chem 16:121620. Van den Berg CMG (1984) Talanta 31:1069621. Zima J, Van den Berg CMG (1994) Anal Chim Acta 289:291622. Muzzarelli RAA, Sipos LG (1971) Talanta 18:853623. Lieberman SH, Zirino H (1974) Anal Chem 46:20624. Johnson KS, Coale KH, Jannasch HW (1992) Anal Chem 64:1070625. Kim YS, Zeitlin H (1973) Separ Sci 7:1626. Shigematsu T, Nishikawa Y, Hiraki K, Goda S, Tsujumatu (1971) Japan Analyst 20:575627. Lee C, Kim NB, Lee IC, Chung KS (1977) Talanta 24:241628. Lieser KH, Calmano W, Heuss E, Neitzert V (1977) J Radioanal Chem 37:717629. Elderfield H, Greaves HJ (1981) Trace metals in seawater. In: Wong CS et al. (eds)

Proceedings of a NATO advanced research institute on trace metals in seawater, Sicily,Italy. Plenum Press, New York, p 427

630. Hooker PJ (1974) BA Thesis. University of Oxford631. Hooker PJ, O’Nions RK, Pankhurst RJ (1975) Chem Geol 16:189632. Korkisch J, Arrhenius G (1964) Anal Chem 36:850633. Faris JP, Warton JW (1962) Anal Chem 34:1077634. Desai HB, Krishnamoorthy IR, Sandar Das M (1964) Talanta 11:1249635. Brunfelt AO, Steinnes E (1969) Analyst 94:979636. Möller P, Dulski P, Luck J (1992) Spectrochim Acta B 47:1379637. Elderfield H, Greaves MJ (1982) Nature 296:214638. Doherty W (1989) Spectrochim Acta B 44:263639. Stroh A (1992) Atom Spectrosc 13:89640. Evans EH, Giglio JJ (1993) J Anal Atom Spectrosc 8:1641. Daidoji H, Tamura S, Matsubara M (1985) Bunseki Kagaku 34:340642. Greaves MJ, Elderfield H, Klinkhammer GP (1989) Anal Chim Acta 218:265643. Piepgras DJ, Jacobsen SB (1992) Geochim Cosmochim Acta 56:1851644. LeBaar HJW, Bacon MP, Brewer PG (1985) Geochim Cosmochim Acta 49:1943645. LeBaar HJW, Brewer PG, Bacon MP (1985) Geochim Cosmochim Acta 49:1961646. Sawatari H, Toda T, Saizuka T, Kimata C, Itoh A, Haraguchi H (1995) Bull Chem Soc

Japan 68:3065647. Esser BK, Volpe A, Kenneally JM, Smith DK (1994) Anal Chem 66:1736648. Halicz L, Gavrieli I, Dorfman E (1996) J Anal Atom Spectrosc 11:811649. Shabani MB, Akagi T, Shimizu H, Masuda A (1990) Anal Chem 62:2709650. Shabani MB, Akagi T, Shimizu H, Masuda A (1992) Anal Chem 64:737651. Murty DSR, Chakrapani G (1996) J Anal Atomic Spectrosc 11:815652. Zhang T-H, Shan X-Q, Liu R-X, Tang H-X, Zhang S-Z (1998) Anal Chem 70:3964653. Mascini M (1971) Anal Chim Acta 56:316


654. Jagner D, Kerstein A (1971) Anal Chim Acta 57:185655. Jagner D, Kerstein A (1972) Anal Abstr 23:1190656. Pausch JB, Margerum DV (1969) Anal Chem 41:226657. Atienza J, Henrero MA, Maquiera A, Duchades R (1992) Crit Rev Anal Chem 64:231658. Tao H, Miyazaki A, Bansho K, Umezaki Y (1984) Anal Chem 56:181659. Armannsson H (1979) Anal Chim Acta 110:21660. Olsen S, Luiz CR, Pessenda JR, Hansen EO (1983) Analyst 108:905661. Fang Z, Ruzicka J, Hansen EH (1984) Anal Chim Acta 164:23662. Cabezon LM, Caballero M, Cela R, Perez-Bustamante JA (1984) Talanta 31:597663. Orren MJ (1971) Journal of South African Chemical Institute (1971) 24:96664. Zhuang Z, Yang C, Wang X, Huang B (1996) Fresen J Anal Chem 355:277665. Chakraborti D, Adams F, Van Mol W, Irgolic KJ (1987) Anal Chim Acta 196:23666. Jin L, Wui D, Ni Z (1987) Huaxue Xuebao 45:808667. Rodienova TV, Ivanov VM (1986) Zh Anal Khim 41:2181668. Allain P, Mauras Y (1984) Anal Chim Acta 165:141669. Doidge PS (1985) Spectrochim Acta 40B:569670. Manning DC, Slavin W, Myers S (1979) Anal Chem 51:2375671. Bengtsson M, Danielsson LG, Magnusson B (1979) Anal Lett 12:A13672. Campbell WC, Ottaway JH (1977) Analyst 102:495673. Stein VB, Canelli G, Richards AH (1980) Int J Environ Anal Chem 8:99674. Cachetti G, Cvelic-Lazzari N, Rolle E (1981) Metodi Analitica per le Aeque 2:1675. Montgomery JR, Peterson GN (1980) Anal Chim Acta 117:397676. Lyons MB, Fitzgerald WF (1981) Trace metals in seawater. In: Wong CS et al. (eds)


677. Brooks RR, Presley BJ, Kaplan IR (1967) Talanta 14:809678. Yates J (1982) Technical report TR181. Determination of trace metals (cadmium,

copper, lead, nickel, and zinc) in filtered saline water samples. Water Research Centre,Medmenham, UK

679. Boyle EA, Edmond JM (1977) Anal Chim Acta 91:1801680. Brugmann L, Danielsson LG, Magnusson B, Westerlund S (1983) Mar Chem 13:327681. Hoenig M, Wollast R (1982) Spectrochim Acta B 37B:399682. Tominaga M, Bansho K, Umezaki Y (1985) Anal Chim Acta 169:171683. Tominaga M, Bansho K, Umezaki Y (1985) Anal Chim Acta 169:176684. Hayase K, Shitashima K, Tsubota H (1986) Talanta 33:754685. Tikhoirova AA, Patini SA, Morozov NP (1976) Zh Anal Khim 31:282686. The Analytical Report, No. 12 (1988) Perkin-Elmer Ltd., Beaconsfield, UK, p 6687. Slavin W, Carnrick GR, Manning DC (1982) Anal Chem 54:621688. Techniques in Graphite Furnace Atomic Absorption Spectrometry (1982) Perkin-

Elmer Ltd., Beaconsfield, UK, Part No. 0993-8150689. Hodge V, Stallard M, Koide M, Goldberg ED (1986) Anal Chem 58:616690. Abollino O, Aceto M, Sacchero G, Sarzanini C, Mentasti E (1995) Anal Chim Acta

305:200691. Cimadevilla EA-C, Wrobel K, Sanz-Medel A (1995) J Anal Atom Spectrosc 10:149692. Hadeishi T, Church DA, McLaughlin RD, Zak BD, Nakamura M, Chang B (1975) Science

187:348693. Koizumi H, Yasuda K (1975) Anal Chem 47:1679694. Stephens R, Ryan DE (1975) Talanta 22:655695. Stephens R, Ryan DE (1975) Talanta 22:659696. Veinot DE, Stephens R (1976) Talanta 23:849

References 305

697. Koizumi H, Yasuda K (1976) Spectrochim Acta 31B:237698. Koizumi H, Yasuda K (1976) Anal Chem 48:1178699. Stephens R, Ryan DE (1978) Talanta 25:435700. Stephens R (1979) Talanta 26:57701. Lawson JB, Grassam E, Ellis DJ, Keir MJ (1976) Analyst 101:315702. Grassam E, Dawson JB, Ellis DJ (1977) Analyst 102:804703. Koizumi H, Yasuda K (1976) Spectrochim Acta B 31B:523704. Koizumi H, Yasuda K, Katayama M (1977) Anal Chem 49:1106705. Pernandez FJ, Bohler W, Beatty NM, Barnett WB (1981) Atom Spectrosc 2:73706. Slavin W, Manning DC, Carnrick GR (1981) Atom Spectros 2:137707. Grabenski Z, Lehmann R, Radzuik B, Voellkopf U (1984) Atom Spectros 5:87708. De Kersabiec AM, Blanc G, Pinta M (1985) Fresen Z Anal Chem 322:731709. Grobenski Z, Lehmann R, Radzuck B, Voellkopf U (1984) The determination of trace

metals in seawater using Zeeman graphite furnace AAS. In: Atomic SpectroscopyApplication Study No. 686 (1984) Papers presented at Pittsburgh Conference, AtlanticCity, NJ, USA

710. Analytical Methods for Graphite Furnace AAS (1984) Perkin-Elmer711. Cabon JY, Le Bihan A (1995) Spectrochim Acta B 50B:1703712. Andreae MOP (1981) Trace metals in seawater. In: Wong CS et al. (eds) Proceedings of


713. Braman RS, Justen LC, Foreback CC (1972) Anal Chem 44:2195714. Braman RS, Foreback CC (1973) Science 182:1247715. Tompkins MA (1977) Environmental–analytical studies of antimony, germanium, and

tin. Thesis. University of South Florida716. Andreae MO, Froelich PN (1981) Anal Chem 53:287717. Fleming HH, Ide RG (1976) Anal Chim Acta 75:8367718. Crecelius EA (1978) Anal Chem 50:826719. Nakashima S, Sturgeon RE, Willie SN, Berman SR (1988) Anal Chim Acta 207:291720. Foreback CC (1973) Some studies on the detection and determination of mercury,

arsenic, and antimony in gas discharges. Thesis. University of South Florida721. Margoshes M, Scribner BF (1959) Spectrochim Acta 15:138722. Korolev VV, Vainshtein EE (1959) Z Anal Chem 15:658723. Owen LW (1961) Appl Spectros 15:150724. Valente SE, Schrenk WG (1970) Appl Spectros 24:197725. Marinkovic M, Vickers EJ (1971) Appl Spectros 25:319726. Chapman JF, Dale LS, Whittem RN (1973) Analyst 98:529727. Merchant P, Veillon C (1974) Anal Chim Acta 70:17728. Murdick DA Jr, Piepmeier EH (1974) Anal Chem 46:678729. Rippetoe WE, Johnson ER, Vickers TI (1975) Anal Chem 47:436730. Winge RK, Fassel VA, Kniseley RN, DeKalb E, Raas WJ (1977) Spectrochim Acta

32B:327731. Nygaard DD (1979) Anal Chem 51:881732. McLeod CW, Otsuki K, Okamoto H, Fuwa K, Haraguchi H (1981) Analyst 106:419733. Mujazaki A, Kimura A, Bansho K, Amezaki Y (1981) Anal Chim Acta 144:213734. Blockaert JAC, Leis F, Lagua K (1981) Talanta 28:745735. Berman SS, McLaren JW, Willie SN (1980) Anal Chem 52:488736. Sturgeon RE, Berman SS, Desaulniers JAH, Mykytiuk AP, McLaren JW, Russell DS

(1980) Anal Chem 52:1585737. Hiraide M, Ito T, Baba M, Kawaguchi H, Mizuike A (1980) Anal Chem 52:804


738. McLaren JW, Mykytiuk AP, Willie SN, Berman SS (1985) Anal Chem 57:2907739. De Oliveira, McLaren E, Berman SS (1983) Anal Chem 55:2047740. Berndt H, Harms U, Sonneborn M (1985) Fresen Z Anal Chem 322:329741. Beauchemin M, Berman SS (1989) Anal Chem 61:1857742. Chong NS, Norton ML, Anderson JL (1990) Anal Chem 62:1043743. Chapple G, Byrne JP (1996) J Anal Atom Spectrosc 11:549744. Alves LC, Allen LA, Houk RS (1993) Anal Chem 65:2468745. Vandecasteele C, Nagels M, Vanhoe NH, Dams R (1988) Anal Chim Acta 211:91746. Stroh A, Voellkopf U (1993) J Anal Atom Spectrosc 8:35747. Field MP, Cullen JT, Sherrell RM (1999) J Anal Atom Spectrosc 14:1425748. Taylor DB, Kingston HM, Nogay DJ, Koller D, Hutton R (1996) J Anal Atom Spectrosc

11:187749. Seubert A, Petzold G, McLaren JW (1995) J Anal Atom Spectrosc 10:371750. Bettinelli M, Spezia S (1995) J Chromatogr A 709:275751. Bloxam MJ, Hill SJ, Worsfold FJ (1993) Anal Proc 30:159752. Nygaard DD, Hill SR (1979) Anal Lett 12:491753. Whitnack GC, Basselli R (1969) Anal Chim Acta 47:267754. Komatsu M, Matsueda T, Kakiyami H (1971) Japan Analyst 20:987755. David J, Redmond JD (1971) Journal of Electro Anal Chem 33:169756. Florence TM (1972) Journal of Electro Anal Chem 35:237757. Rojohn T (1972) Anal Chim Acta 62:438758. Seitz WR, Jones R, Klatt LN, Mason WD (1973) Anal Chem 45:810759. Clem RG, Litton G, Ornelas LD (1973) Anal Chem 45:1306760. Clem RG (1973) MPJ Application Notes 8:1761. Bubic S, Sipos L, Branica M (1973) Thalass Jugoslavia 9:55762. ChauYK,GachterR,Lum-Shue-ChanK(1974) Journal of theFisheriesResearchBoard,

Canada 31:1515763. Nurnberg HW, Valenta P, Mart L, Raspor B, Sipos L (1976) Fresen Z Anal Chem 282:357764. Florence TM, Batley GE (1977) Talanta 24:151765. Whitnak GC, Sanelli R (1969) Anal Chim Acta 47:269766. Smith JD, Redmond JD (1971) Journal of Electro Anal Chem 33:169767. Sinko I, Dolczal J (1970) Journal of Electro Anal Chem 25:299768. Cambon JP, Cauw G, Monaco A (1977) Review of International Oceanographic

Medicine T. XIVIII, 2:73769. Welghe N, Claeys A (1972) Journal of Electro Anal Chem 35:229770. Brugmann L (1974) Acta Hydrochim Hydrobiol 2:123771. Lund W, Salberg M (1975) Anal Chim Acta 76:131772. Gardner J, Stiff MJ (1975) Water Res 9:517773. Abdullah MI, Reusch Berg B, Klimek R (1976) Anal Chim Acta 84:307774. Duyckaerts G, Gillain G (1977) Essays in memory of Anders Ringbom. Pergamon

Press, Oxford775. Brezonik PL (1974) In: Rubin AJ (ed) Aqueous environmental chemistry of metals.

Ann Arbor Science Publishers, Ann Arbor, Michigan, p 167776. Elder JF (1975) Limnol Oceanogr 20:96777. O’Shea TA, Maney KH (1976) Anal Chem 48:1603778. Batley GE, Matousek JP (1980) Anal Chem 52:1570779. Young J, Gurtisen JM, Apts CW, Crecelius EA (1979) Mar Environ Res 2:265780. Batley GE (1981) Anal Chim Acta 124:121781. Scarponi G, Capodaglio G, Cescon P, Cosma B, Frache R (1982) Anal Chim Acta 135:263782. Brügman L, Danielsson LG, Magnusson B, Westerlund S (1983) Mar Chem 13:329

References 307

783. Clem RG, Hodgson AT (1978) Anal Chem 50:102784. Brügman L (1984) Sci Total Environ 37:41785. Bruland KW, Coale KH, Mart L (1985) Mar Chem 17:285786. Achterberg EP, Van den Berg CMG (1996) Mar Pollut Bull 32:471787. Brihaye C, Gillain G, Duyckaerts G (1983) Anal Chim Acta 148:51788. Cuculic V, Branica M (1996) Analyst 121:1127789. Sipos L, Golimoivski J, Valenta P, Nurnberg HW (1979) Fresen Z Anal Chem 298:1790. Bott A (1995) Current Separation 14:24791. Bond AM, Heritage ID, Thormann W (1986) Anal Chem 58:1063792. Nurnberg HW (1984) Sci Total Environ 37:9793. Mart L, Nurnberg WH, Dryssen D (1981) Trace metals in seawater. In: Wong CS et al.

(eds) Proceedings of a NATO advanced research institute on trace metals in seawater,Sicily, Italy. Plenum Press, New York

794. Valenta P, Mart L, Rutzel H (1977) Journal of Electro Anal Chem 82:327795. Pihlar B, Valenta P, Nurnberg HW (1981) Fresen Z Anal Chem 307:337796. Pietrowicz SR, Springer-Young M, Puig JA, Spencer MJ (1982) Anal Chem 54:1367797. Lund W, Onshus D (1976) Anal Chim Acta 86:109798. Van den Berg CMG (1984) Geochim Cosmochim Acta 48:2613799. Kraznaric D (1984) Mar Chem 15:117800. Andruzzi R, Trazza A, Marrosu G (1982) Talanta 29:751801. Andruzzi R, Trazza A, Marrosu G (1982) Anal Lett (London) 15:1965802. Jagner D, Josefson M, Westerlund S (1981) Anal Chim Acta 129:153803. Drabek I, Pfeiffer O, Madsen P, Sorensen J (1987) Int J Environ Anal Chem 15:153804. Donat JR, Bruland KW (1993) Anal Chim Acta 254:547805. Perez-Pina J, Hernandez-Brito JJ, Herrera-Melian JA, Collado-Sanchez C, Van den Berg

CMG (1994) Electroanalysis (NY) 6:1069806. Van den Berg CMG (1994) J Anal Atom Spectrosc 6:1069807. Van den Berg CMG (1986) Sci Total Environ 49:89808. Eskilsson H, Haroldsson C, Jagner D (1985) Anal Chim Acta 175:79809. Prange A, Knochel A, Michaelis W (1985) Anal Chim Acta 172:79810. Prange A, Knoth J, Stossel R-P, Bodekker H, Kramer (1987) Anal Chim Acta 195:275811. Haarich M, Schmidt D, Freimann P, Jacobsen A (1993) Spectrochim Acta B 48:183812. Knochel A, Prange A (1981) Fresen Z Anal Chem 306:252813. Prange A, Omatsu M, Muskimoto S (1984) X-Ray Spectrometry 13:83814. Morris AW (1968) Anal Chim Acta 42:397815. Robertson DE, Carpenter R (1972) Report No. BNWL-SA-4455 Access 22569. National

Technical Information Service, Springfield, VA, USA816. Robertson DR, Carpenter R (1974) Report No. NAS-NS-3114 Access 26418. National

Technical Information Service, Springfield, VA, USA817. Piper DZ, Goles GG (1969) Anal Chim Acta 47:560818. Ghoda S (1972) Bull Chem Soc Japan 45:1704819. Brine E (1972) Report Aktiebolaget Atomenergi AE-466820. Lo JM, Wei JC, Yeh SJ (1977) Anal Chem 49:1146821. Greenberg RR, Kingston HM (1982) J Radioanal Chem 71:147822. Greenberg RR, Kingston HM (1983) Anal Chem 55:1160823. Murthy RSS, Ryan DE (1983) Anal Chem 55:682824. Stiller M, Mantel M, Rappaport MS (1984) J Radioanal Nucl Chem 83:345825. Holzbecker J, Ryan DE (1982) J Radioanal Chem 74:25826. Chow TJ, Goldberg ED (1981) Journal of Mar Res 20:163827. Chow TJ, Patterson CC (1966) Earth Planet Sci Lett 1:397


828. Wilson JD, Webster RK, Milner GWC, Barnett GA, Smales AA (1960) Anal Chim Acta23:505

829. Leipziger FD (1965) Anal Chem 37:171830. Paulsen P, Alvarez R, Kelleher D (1969) Spectrochim Acta B 24:535831. Chow TJ (1968) Journal of the Water Pollution Control Federation 399832. Patterson CC, Settle DM, Glover B (1976) Mar Chem 4:305833. Schalue B, Patterson CC (1980) The Occurrence of lead in the northeastern pacific,

and the effects of anthropogenic inputs. In: Brancia M, Konrad Z (eds) Lead in themarine environment. Pergamon Press, Oxford, pp 31–43

834. Stukas VJ, Wong CS (1981) Science 211:1424835. Murozumi M (1981) Isotope dilution mass spectrometry of copper, cadmium, thal-

lium, and lead in marine environments. Presented at The American Chemical SocietyCongress, Hawaii

836. Stuckas VJ, Wong CS (1981) Trace metals in seawater. In: Wong CS et al. (eds) Proceed-ings of a NATO advanced research institute on trace metals in seawater, Sicily, Italy.Plenum Press, New York, p 513

837. Wu J, Boyle EA (1997) Anal Chem 69:2464838. Riel G (1995) In: King SE (ed) Proceedings of the workshop on monitoring nuclear

contamination in arctic seas. Naval Research Laboratory, Washington, DC839. Nagaosa Y, Kawabe H, Bond AM (1991) Anal Chem 63:28840. Shijo Y, Sato H, Uehara N, Aratake S (1996) Analyst 121:325841. Riccardo AA, Muzzarelli RAA, Raith G, Tubertini O (1970) J Chromatogr 47:414842. Bloxham MJ, Hill SJ, Worsfold FJ (1993) Anal Proc 30:160843. Morel FMM, Westall JC, O’Melia CR, Morgan JJ (1975) Environ Sci Tech 9:756844. Galloway JN (1979) Geochim Cosmochim Acta 43:207845. Long DT, Angino EE (1977) Geochim Cosmochim Acta 4:1183846. Dyrssen D, Wedborg M (1980) Major and minor elements, chemical speciation in

estuarine waters. In: Olausson E, Cato I (eds) Chemistry and biogeochemistry ofestuaries. Wiley, New York, pp 71–119

847. Amdurer M (1981) Trace metals in seawater. In: Wong CS et al. (eds) Proceedings ofa NATO advanced research institute on trace metals in seawater, Sicily, Italy. PlenumPress, New York, p 537

848. Fukai R, Huynh-Ngoc L (1975) Journal of Oceanographic Society of Japan 31:1849. Young DR, Jan T-K, Hershelman GP (1980) Cycling of zinc in the nearshore marine

environment. In: Nriagu JO (ed) Zinc in the environment. Wiley, New York, pp 298–335850. Jennings CD (1978) Mar Sci Comm 4:49851. Cranston RE, Murray JW (1980) Limnol Oceanog 25:1106852. Van den Berg CMG (1983) Anal Proc 20:455853. Midorikawa M, Tanoue E, Sugimura Y (1990) Anal Chem 62:1737854. Mackey DJ (1982) Mar Chem 11:169855. Latouche C, Dumon LC, Lavaux G, Pedemaj P (1993) Int J Environ Anal Chem 51:177856. Baskaran M, Murphy DJ, Santschi PH, Orr JC, Schink DR (1993) Deep-Sea Res 40:849857. Magni EG, Tiziana G, Frache R (1995) Anal Proc including Anal Comm 32:267858. Zhang H, Davison W, Grime GW (1995) ASTM Special Technical Publication STP 1293,

p 17859. Resing JA, Measures CI (1994) Anal Chem 66:4105860. Sturgeon RE, Willie SN, Berman SS (1985) Anal Chem 57:6861. Pierce FD, Brown HR (1977) Anal Chem 49:1417862. Yusof AM, Ikhson ZB, Wood AKH (1994) J Radioanal Nucl Chem 179:277863. Sperling KR (1982) Fresen Z Anal Chem 310:254

References 309

864. Ganzerli VMT, Stella R, Maggi L, Cicerci G (1987) J Radioanal Nucl Chem 114:105865. Janzer VJ (1973) Journal of Research of US Geological Society 1:113866. Zhang HT, Holzenbecker J, Ryan D (1983) Anal Chim Acta 149:385867. Henman KG (1980) Toxicol Environ Chem Rev 3:111868. Marshall MA, Mottola HA (1985) Anal Chem 57:729869. Kiriyama T (1993) Bunseki Kagaku 42:223870. Kagoya S, Ueda J (1994) Bull Chem Soc Japan 67:1965871. Reimer RA, Miyazaki A (1992) J Anal Atom Spectrosc 7:1239872. Miyazaki A, Reimer RA (1993) J Anal Atom Spectrosc 8:449873. Liu ZS, Huang SD (1995) Spectrochim Acta B 50:197874. Flynn WW (1973) Anal Chim Acta 67:129875. Billah M, Honjo T, Terada K (1993) Anal Sci 9:251876. Olafsson J (1981) Trace metals in seawater. In: Wong CS et al. (eds) Proceedings of

a NATO advanced research institute on trace metals in seawater, Sicily, Italy. PlenumPress, New York, p 476

877. Kwokal Z, May K, Branica M (1994) Sci Total Environ 154:63878. Fujita M, Iwashima K (1981) Environ Sci Tech 15:929879. Lo JM, Yu JC, Hutchison F, Wai CM (1982) Anal Chem 54:2536880. Fernandez-Garcia M, Pereiro-Garcia R, Bordel-Garcia N, Sanz-Medel A (1994) Talanta

41:183881. Chan KM, Riley JP (1966) Anal Chim Acta 36:220882. Ohta N, Fujita M, Tomura K (1979) Bunseki Kagaku 28:277883. Prabhu VG, Zarapkar LR, Das MS (1980) Microchim Acta 11:67884. Kim YS, Zeitlin H (1969) Anal Chim Acta 46:1885. Head P, Burton JD (1970) J Mar Biol Ass UK 50:439886. Butler LRP, Matthews PM (1966) Anal Chim Acta 36:319887. Chau YK, Lum Shue Chan K (1969) Anal Chim Acta 48:205888. Akama Y, Nakai T, Kawamura F (1979) Nippon Kaisui Gakkai-shi 33:180889. Muzzarelli RAA (1971) Anal Chim Acta 54:133890. Vanderborght BM, Van Grieken RE (1977) Anal Chem 49:311891. Yoshimura K, Hiraoka S, Tarutani T (1982) Anal Chim Acta 142:101892. Riley JP, Taylor D (1968) Anal Chem 41:175893. Ficklin WH (1982) Anal Lett 15:865894. Riley JP, Taylor D (1968) Anal Chim Acta 40:479895. Van der Sloot HA, Wals GD, Das HA (1977) Anal Chim Acta 90:193896. Hkagi T, Noriji Y, Matsui M, Haraguchi H (1985) Appl Spectros 39:662897. Zlatkis A, Bruening W, Bayer E (1971) Anal Chim Acta 56:399898. Bruland KW (1980) Science Letters 47:177899. Rampon H, Cuvelier R (1972) Anal Chim Acta 60:226900. Lee DS (1982) Anal Chem 54:1182901. Cowen JP, Hodge VF, Folson TR (1977) Anal Chem 49:494902. Reid DF, Key RM, Schink DR (1979) Earth Planet Sci 43:223903. Chen Q, Aarkrog A, Nielson SP (1993) J Radioanal Nucl Chem 172:281904. Bomber AM, Equillor HE, Oliveira AA (1994) J Radioanal Nucl Chem 177:27905. Hirose K, Sugimura YJ (1985) J Radioanal Nucl Chem 92:363906. Sturgeon RE, Willie SN, Berman SS (1986) Anal Chem 58:1140907. Siu KWH, Berman SS (1984) Anal Chem 56:1806908. Avila AK, Curtius AJ (1994) J Anal Atom Spectrosc 9:543909. Honjo T, Nakata T (1987) Bull Chem Soc Japan 60:2271910. Chen Q, Dahlgaard H, Nielson SP (1994) Anal Chim Acta 285:177


911. Momoshima N, Sayad M, Takashima Y (1993) Radioanalytica Acta 63:73912. Tao H, Rajendran RB, Quetel CR, Tominaga M, Nakazato T, Miyazaki A (1999) Anal

Chem 71:4208913. Korkisch J, Koch W (1973) Microchim Acta 1:157914. Korkisch J (1972) Microchim Acta 6:87915. Kuroda R, Oguma K, Mukai N, Iwamoto M (1987) Talanta 34:433916. Wu TL, Lambert L, Hastings D, Banning D (1980) Bull Environ Contam Toxicol 24:411917. Hashimoto T (1971) Anal Chim Acta 56:347918. Kim YS, Zeitlin H (1971) Anal Chem 43:1390919. Nishimura M, Matsunaga K, Kudo T, Obara F (1973) Anal Chim Acta 65:446920. Dupont V, Auger Y, Jeandel C, Wartel M (1991) Anal Chem 63:520921. Nowicki JL, Private Communication922. Sastry VN, Frishnamoorthy TM, Sarma TP (1968) Current Science (Bombay) 38:279923. Tsalev DL, Alimarin TP, Neiman SI (1972) Zh Anal Khim 27:1223924. El-Enamy FF, Mahmond KF, Varma MM (1979) Journal of the Water Pollution Control

Federation 51:2545925. Pellenberg RE, Church TM (1978) Anal Chim Acta 97:81926. Jan TK, Young DR (1978) Anal Chem 50:1250927. Danielsson LG, Magnusson B, Westerlund S (1982) Anal Chim Acta 144:183928. Van Geen A, Boyle E (1990) Anal Chem 62:1705929. Bafti F, Cardinale AM, Bruzzone R (1992) Anal Chim Acta 270:79930. Ogura H, Oguma K (1994) Microchem J 49:220931. Yang Y, Huang W, Liu D (1987) Xiamen Daxue Xuebao Ziran Kexueban 26:96932. Dai SE, Chen TC, Wong GTF (1990) Anal Chem 62:774933. Buchanan AS, Hannaker P (1984) Anal Chem 56:1379934. Pai S-C, Chen T-C, Wong GTF, Hung CC (1990) Anal Chem 62:774935. Nakashima S, Sturgeon RE, Willie SN, Berman SS (1988) Fresen J Anal Chem 330:592936. Volkan M, Ataman OY, Howard AG (1987) Analyst 112:1409937. Terada K, Morimoto K, Kiba T (1980) Anal Chim Acta 116:127938. Lau CR, Yang M (1994) Anal Chim Acta 287:111939. Kawamura S, Sadao S, Katsumi K (1971) Anal Chim Acta 56:405940. Johansson M, Emteborg H, Glad B, Reinholdsson F, Baxter DC (1995) Fresen J Anal

Chem 351:461941. Blain S, Appriou P, Handel H (1993) Anal Chim Acta 272:91942. Orians KJ, Boyle EA (1993) Anal Chim Acta 282:63943. Hirayama K, Unchara NN (1987) Daigaku Kogakubu Kiyo Bunrui K 28:149944. Rao RR, Chat A (1993) J Radioanal Nucl Chem 168:439945. Civici N (1994) J Radioanal Nucl Chem 186:303946. Zeng H-H, Thompson RB, Maliwal BP, Fores GR, Moffett JW, Flecke CA (2002) Anal

Chem 74:6807947. Nakazato T, Tao H, Toniguchi T, Isshiki K (2002) Talanta 58:121948. Izumi Y, Kiyotaki F, Minato T, Seida Y (2002) Anal Chem 74:3819949. Baena JR, Gallego M, Vacarcel M (2002) Anal Chem 74:1519950. Wen B, Shan X, Xu SG (1999) Analyst 124:621951. Tony KA, Kartikeyan S, Vijayalakshmy B, Rao Tp, Iyer CSP (1999) Analyst 124:191952. Chang HJ, Sung YH, Huang SD (1999) Analyst 124:1695953. Gee Ak, Bruland KW (2002) Geochim Cosmochim Acta 66:3063954. Labonne M, Ottman DB, Luck JM (2002) Applied Geochemistry 17:1351955. Liu HW, Jiang SJ, Liu SH (1999) Spectrochim Acta B 54:1367

References 311

956. Warnken KW, Gill GA, Wen L-S, Griffin LL, Santschi PH (1999) J Anal AtomicSpecroscopy 14:247

957. Wen B, Shan XQ, Liu RX, Tang HY (1999) Fresen J Anal Chem 363:251958. Esser B, Volpe AM (2002) Mar Chem 79:67959. Esser B, Volpe AM (2002) Mar Chem 36:2836960. Worsfold PJ, Achterberg EP, Bowie AR, Sandford R, Montoura RF (2000) Ocean Science

and Technology 1:71961. Achterberg EP, Braungardt C, Whitworth DJ (2000) Ocean Science and Technology

1:277962. Achtenberg EP (2000) Int J Environ Pollution 13:249963. Bermego Barrera P, Barciela Alonso MC, Ferron Novais M, Bermego Barrera A (1996)

Fresen Z Anal Chem 355:174964. Hirao Y, Fukumoto K, Sugisaki H, Kuimura K (1979) Anal Chem 51:651

6 Cations in Estuary, Bay, and Coastal Waters

6.1Ammonium

Berg and Abullah [1] and Mantoura and Woodward [2] have described in-dophenol blue methods for the automated determination of ammonia in estu-arine waters.

The reaction manifold describing the automated determination of ammoniais shown in Fig. 6.1. Two alternative modes of sampling are shown: discrete andcontinuous. Discrete 5 ml samples contained in ashed (450 ◦C) glass vials aresampled from an autosampler (Hook and Tucker model A40-11; 1.5 min sam-ple/wash). For high-resolution work in the estuary, the continuous samplingmode is preferred. The indophenol blue complex was measured at 630 nm witha colorimeter and the absorbance recorded on a chart recorder.

Mantoura and Woodward [2] overcame the problem of magnesium precipi-tation by ensuring a stoichiometric excess of citrate (about 120%). These work-ers believe that “salt errors” occurring with estuarine samples originate frompoor pH buffering rather than ionic strength variations. They, in fact, used phe-nol at a concentration of 0.06 M to make the system self-buffering. Even in thepresence of 1 mg NH3–N per litre the indophenol blue reaction will consumeonly 3% of the phenol leaving most of the phenol to act as a pH buffer. Ethanolwas used to solubilise the high concentration of phenol used in the system. Thesalt error of this method, as determined by standard addition of ammonia intowaters of different salinities, is shown in Fig. 6.2(a). When compared with othermethods, the method displays minimal salt error (about 8%) even though thefinal pH of the river water mixture (pH 10.9) was greater than seawater (pH 9.9).

In addition to the chemical effects of varying salinity, there are opticalinterferences in colorimetric analysis which are peculiar to estuarine samples.Saline waters and river waters have, in the absence of colorimetric reagents, anapparent absorbance arising from:

1. Refractive bending of light beams by sea salts-“refractive index blank” [3]2. Background absorbance by dissolved organics of riverine origin.

The former is a function of the optical geometry of the light beam and the flowcell, and the latter is related to the organic loading of river water. Figure 6.2(b)

314 6 Cations in Estuary, Bay, and Coastal Waters

Figure 6.1. Manifold for the automatic determination of ammonia (# on pump tube R1indicates Solvaflex tubing). Source: [2]

shows that both are linearly related to salinity, which makes optical blankcorrections easy to apply to estuarine samples.

6.2Arsenic

6.2.1Hydride Generation Atomic Spectrometry

Amankwah and Fasching [4] have discussed the determination of arsenic (V)and arsenic (III) in estuary water by solvent extraction and atomic absorptionspectrometry using the hydride generation technique.

6.3Barium


Epstein and Zander [5] determined barium directly in estuarine and sea waterby graphite furnace atomic absorption spectrometry.

6.4 Cadmium 315

Figure 6.2. (a) The effects of salinity on the sensitivity of standard additions of ammoniain laboratory mixed waters (•) and in waters from the Tamar estuary (�) expressed aspercentageof response in riverwater. Forcomparison, the salt errorcurves reportedbyLoderand Gilbert [3] are also shown (. . . and - - -, respectively). (b) Contribution of reactive indexand organic absorbance to the optical blacks in the Chemlab Colorimeter. • = River water–seawater mixture. ◦ = De-ionized water–seawater mixture. Source: [2]

6.4Cadmium


Gardner [6] has reported a detailed statistical study involving ten laboratoriesof the determination of cadmium in coastal and estuarine waters by atomic ab-sorption spectrometry. The maximum tolerable error was defined as 0.1 µg/lor 20% of sample concentration, whichever is the larger. Many laboratoriesparticipating in this work did not achieve the required accuracy for the deter-mination of cadmium in coastal and estuarine water. Failure to meet targets isattributable to both random and systematic errors.


6.5Calcium and Magnesium

Titration

Arey et al. [35] have described a method for the determination of calcium andmagnesium ions in estuarine waters. Calcium is determined by titration witha chelating agent ethylene bis oxyethylene nitrilo/tetra acetic acid and calciumand magnesium together with EDTA. A cupric ion-selective electrode is usedfor both determinations and the indicator is ethylene bis (oxyethylene-nitrilo)tetra-acetic acid cupric chelate for calcium alone and EDTA cupric chelate forthe two ions together; the ratioof ammoniumtoassociatedammoniumhydrox-ide in the buffer solution must be carefully adjusted for each titration to give theproper cupric ion concentration. The method is suitable for routine analysisbut gives consistently lower results than atomic absorption spectrometry.

The lower limit of detection of this method is 0.2 mg/l for calcium andmagnesium.

6.6Copper

6.6.1Titration Procedure

Berger et al. [7] applied a fluorescence quenching titration method to the mea-surement of the complexation of copper (II) in the Gironde Estuary. Relativelyhigh values of residual fluorescence after titration indicated that much organicfluorescing material does not bind to divalent copper. The titration was per-formed in a flow-through system thermostated at 25 ◦C under nitrogen. ThepH is adjusted for each step at 8.0 with 0.01 M potassium hydroxide or nitricacid (pH 8.0 is very close to the natural pH of the Gironde waters). After theaddition of copper (II) solution, the sample is circulated through the cuvette forseveral minutes before measurement. An increase in Rayleigh scattering, whichwas measured along with fluorescence, signifies aggregation. When Rayleighscattering doubled its original value before copper (II) was added, the titrationwas stopped.


Nelson [8] studied voltammetric measurement of copper (II)-organic interac-tions in estuarine waters. Based on results of previous studies on the effectsof organic matter on adsorption of copper at mercury surfaces, Nelson de-veloped a method to evaluate the interactions between divalent copper and

6.7 Mercury 317

organic ligands, based on ligand exchange. The copper/organic species com-peted with glycine, which formed copper glycinate and those two complexescould be distinguished voltammetrically, since copper glycinate gave a highersurface excess of copper at a gelatin-coated hanging-drop mercury electrode.The method was applied successfully to both chloride media and estuarinewaters. It was demonstrated that estuarine waters contained two types of lig-and capable of binding divalent copper; humic material with polyelectrolytetype binding, and discrete ligands, with stability constants of about 109. Theextent of binding by humic material decreased down the estuary as a result ofdilution and increased salinity.

Nelson [9] also examined the role of organic matter in the uptake of copperat a mercury electrode. Experiments were carried out on the induced adsorp-tion of copper on a hanging-drop mercury electrode in a stirred solution, usingchloride media with added complexing ligand and organic surfactant, and inestuarine water containing added surfactant (gelatin). Copper chloride wasthe most important copper species adsorbed on the electrode, and adsorptionwas enhanced by the presence of organic films, which could provide a criticalpathway for reducing, divalent copper in estuarine waters. The compositionof organic monolayers could be determined by utilising adsorption of diva-lent and monovalent copper as electroactive probes and determining solutioncopper-organic binding.

Shuman and Michael [10] applied a rotating disk electrode to the measure-ment of copper complex dissociation rate constants in marine coastal waters.An operational definition for labile and non-labile metal complexes was estab-lished on kinetic criteria. Samples collected off the mid-Atlantic coast of USAshowed varying degrees of copper chelation. It is suggested that the techniqueshould be useful for metal toxicity studies because of its ability to measureboth equilibrium concentrations and kinetic availability of soluble metal.

Wang et al. [11] used a remote electrode operated in the potentiometricstripping mode, for the continous boatside determination of copper distribu-tion patterns in San Diego Bay (CA, USA).

6.7Mercury

6.7.1Miscellaneous

Bio-assay methods have been used to obtain estimates of low mercury concen-trations (5–20 µg/l) in seawater (Davies and Pirie [13]). This method is usefulfor detecting comparatively small enhancements over background mercuryconcentrations in estuarine and sea water.

This method consists of suspending 70 mussels (Mytilus edulis) each ofa standard weight, for a standard time in a plastic coated wire cage 2 m below


the surface. Mercury in the mussels was determined by cold vapour atomicabsorption spectrometry [14, 15]. The procedure is calibrated by plotting thedetermined mercury content of mussels against the mercury content of the seawater in the same area.

6.8Manganese

6.8.1Polarography

Knox and Turner [12] have described a polarographic method for man-ganese (II) in estuarine waters which covers the lower concentration range10–300 µg/l. The method, which is specific to manganese (II) and its labilecomplexes, is used in conjunction with a colorimetric technique to comparethe levels of manganese (II) and total dissolved manganese in an estuarinesystem. They showed that polarographically determined manganese (II) canvary widely from 100% to less than 10% of the total dissolved manganese, de-termined spectrophotometrically at 450 nm by the formaldoxine method [16]calibrated in saline medium to overcome any salt effects. It is suggested that themanganese not measured by the polarographic method is in colloidal form.

6.9Selenium

6.9.1Hydride Generation Graphite Furnace Atomic Absorption Spectrometry

Willie et al. [17] used the hydride generation graphite furnace atomic absorp-tion spectrometry technique to determine selenium in saline estuary watersand sea waters. A Pyrex cell was used to generate selenium hydride whichwas carried to a quartz tube and then a preheated furnace operated at 400 ◦C.Pyrolytic graphite tubes were used. Selenium could be determined down to20 ng/l. No interference was found due to, iron copper, nickel, or arsenic.

Cutter [18] has studied the application of the hydride generation method tothe determination of selenium in saline waters.

6.10Tin


Tributyltin has been determined in estuarine waters by high-performanceliquid chromatography with fluorometric detection in a method described


by Ebdon and Alonso [19]. The Bu3Sn+ is quantitatively retained from 100–500 mL of sample on a 4 cm long ODS column. The ODS column was back-flushed with methanol-water containing ammonium acetate onto a PartisilSCX analytical column. The eluent was mixed with acetic acid, Morin, andTriton X-100 for fluorometric detection.

6.11Multication Analysis

6.11.1Heavy Metals, Isotope Dilution, Spark Source Mass Spectrometry,and Inductively Coupled Plasma Atomic Emission Spectrometry

The determination of trace elements in coastal and seawater is pursued withgreat difficulty [20]. Quantitation of extremely low concentrations of ana-lyte (0.02–10 µg/l) accompanied by a matrix consisting of 3.5% dissolvedsolids in sea water imposes great demands on instrumental techniques. Sam-ple preparation schemes designed to both preconcentrate the trace elementsand separate them from major interfering components prior to analysis arenumerous. All such methods invariably increase sample manipulation andthe relatively large amounts of reagents and container surfaces brought intocontact with the sample often give rise to unacceptably high and/or randomprocedural blanks. These problems are exacerbated by a lack of standardreference materials which would permit detection of systematic errors suchas contamination or analyte losses introduced during sample manipulationand the presence of matrix or spectral interferences perturbing instrumentresponse.

A logical approach which serves to minimise such uncertainties is theuse of a number of distinctly different analytical methods for the determi-nation of each analyte wherein none of the methods would be expected tosuffer identical interferences. In this manner, any correspondence observedbetween the results of different methods implies that a reliable estimate ofthe true value for the analyte concentration in the sample has been obtained.To this end Sturgeon et al. [21] carried out the analysis of coastal seawaterfor the above elements using isotope dilution spark source mass spectrom-etry. GFA-AS, and ICP-ES following trace metal separation-preconcentration(using ion exchange and chelation–solvent extraction), and direct analysisby GFA-AS. These workers discuss analytical advantages inherent in such anapproach.

Table 6.1 shows the results obtained by Sturgeon et al. [21] for a storedcoastal water sample. The mean concentrations and standard deviations ofreplicates (after rejection of outliers on the basis of a simple c test-function)are given for each method of analysis. Each mean reflects the result of four ormore separate determinations by the indicated method [21].


Table 6.1. Analysis of seawater sample A [21]

Concentration, µg/l

GFAAS ICPES IDSSMS

Element Direct Chelation– Ion exchange Ion exchangeextraction

Fe 1.6 ± 0.2a 1.5 ± 0.1 1.5 ± 0.6 1.4 ± 0.1Mn 1.6 ± 0.1 1.4 ± 0.2 1.5 ± 0.1 NDb

Cd 0.20 ± 0.04 0.24 ± 0.04 ND 0.28 ± 0.02Zn 1.7 ± 0.2 1.9 ± 0.2 1.5 ± 0.4 1.6 ± 0.1Cu ND 0.6 ± 0.2 0.7 ± 0.2 0.7 ± 0.1Ni ND 0.33 ± 0.08 0.4 ± 0.1 0.37 ± 0.02Pb ND 0.22 ± 0.04 ND 0.35 ± 0.03Co ND 0.018 ± 0.008c ND 0.020 ± 0.003d

a Precision expressed as standard deviationb ND: Not determinedc Preconcentrated hundredfold by ion exchanged Spark source mass spectrometry – internal standard method

Sturgeon et al. [21, 22] conclude that direct analysis by GFA-AS is a fast,accurate method for the determination of iron, manganese, zinc, and cad-mium in seawater when these elements are present at concentrations above0.2, 0.2, 0.4 and 0.01 µg/l, respectively (offshore seawater concentrations ofthese metals are: iron ∼0.5 µg/l, manganese ∼0.05 µg/l, zinc ∼0.2 µg/l, andcadmium ∼0.05 µg/l). Below these levels, and for the other elements stud-ied, chelation–solvent extraction using ammonium pyrrolidine oxine diethyl-dithiocarbamate/methyl isobutylketone incombinationwithaback-extractioninto an acidic aqueous phase prior to determination by GFA-AS is the mostuseful technique for multielement determinations when small volumes of sea-water are available (e.g., 50 × preconcentration on 100 mL aliquot of seawater).If significantly greater preconcentration is required, or if greater volumes ofpreconcentrate solution are needed as, for example, when analysis is completedby ICP-ES, the more laborious method of ion exchange (using Chelex 100) maybe used.

Although ICP-ES is a multielement technique, its inferior detection limits(relative to GFA-AS) necessitate the processing of relatively large volumes ofseawater. 250 mL aliquots were found to be useful for the analysis of iron, man-ganese, zinc, copper, and nickel. Extension of the method to include cadmium,cobalt, chromium, and lead would require improvements in the preconcentra-tion procedure.

Apte and Gunn [23] used liquid–liquid extraction, involving 1:1:1 trichlor-ethane extraction of the ammonium pyrrolidine dithiocarbamates to concen-trate copper, nickel, lead, and cadmium from estuary water. (Detection limits


achieved using electrothermal atomic absorption spectrometry were, respec-tively, 0.3, 0.02, 0.7, and 0.5 µg/l).

Stein et al. [24] have described a simplified, and rapid GFA-AS method fordetermining low concentration of cadmium, lead, and chromium in estuarinewaters. To minimise matrix interferences, nitric acid and ammonium nitrateare added for cadmium and lead; nitric acid only is added for chromium.Then, 10, 20 or 50 µl of the sample or standard (the amount depending on thesensitivity required) is injected into a heated graphite atomiser, and specificatomic absorbance is measured. Analyte concentrations are calculated fromcalibration curves for standard solutions in demineralised water for chromiumor an artificial sea water medium for lead and cadmium.

Detection limits (µg/l), were 0.1 for cadmium, 4 for lead, and 0.2 forchromium. For cadmium (0.5 and 5 µg/l,), lead (4 and 50 µg/l), and chromium(1 and 10 µg/l) in half-strength artificial seawater, the relative standard devia-tions (n = 10) were 20 and 0.5, 18 and 10.4, and 25.0 and 8.0%, respectively.

Pellenberg and Church [25] sampled stored and processed saline water sam-ples from the Delaware Bay estuary in a variety of ways to allow different meth-ods of maintaining their integrity to be compared. Samples were processedonboard ship, immediately after collection, by extraction with ammoniumpyrollidinedithiocarbamate in methyl isobutyl ketone.

Duplicate samples were processed onshore after a variety of storage proce-dures. All samples were analysed for copper and iron by GFA-AS. Only samplesfiltered (< 1 µm), acidified, and stored frozen gave extractable copper and ironresults comparable with those for samples extracted immediately after col-lection. Cold storage with sample acidification in polyethylene containers ap-peared less satisfactory. Organic extracts from samples processed onboard arebest retained in all-Teflon containers pending complete digestion and analysisonshore. Unless clean (ultra-filtered air) conditions can be ensured onboard,the estuarine water samples are best returned in a filtered, acidified, and frozencondition for onshore processing.

Gardner and Yates [26] developed a method for the determination of totaldissolved cadmium and lead in estuarine waters. Factors leading to the choiceof a method employing extraction by chelating resin, and analysis by carbonfurnace atomic absorption spectrometry, are described. To ensure completeextraction of trace metals, inert complexes with humic-like material are de-composed by ozone [27]. The effect of pH on extraction by and elution fromchelating resin is discussed, and details of the method were presented. Theseworkers found that at pH 7 only 1–2 minutes treatment with ozone was neededto completely destroy complexing agents such as EDTA and humic acid in thesamples.

Kingston et al. [32] preconcentrated the eight transition elements cadmium,cobalt, copper, iron, manganese, nickel, lead, and zinc from estuarine andseawater using solvent extraction/chelation and determined them at sub ng/llevels by GFA-AS.


Yamamoto et al. [33] have studied the differential determination of heavymetals according to their oxidation states by flameless atomic absorption spec-trometrycombinedwithsolventextractionwithammoniumpyrrolidinedithio-carbamate or sodium diethyldithio-carbamate.


Heavy Metals

Batley [28] examined the techniques available for the in situ electrodeposi-tion of lead and cadmium in estuary water. These included anodic strippingvoltammetry at a glass carbon thin film electrode and the hanging drop mer-cury electrode in the presence of oxygen and in situ electrodeposition onmercury coated graphite tubes. Batley [28] found that in situ deposition oflead and cadmium on a mercury coated tube was the more versatile tech-nique. The mercury film, deposited in the laboratory, is stable on the driedtubes which are used later for field electrodeposition. The deposited met-als were then determined by electrothermal atomic absorption spectrometry,Hasle and Abdullah [29] used differential pulse anodic stripping voltammetryin speciation studies on dissolved copper, lead, and cadmium in coastal seawater.

Clem and Hodgson [27] discuss the temporal release of traces of cadmiumand lead in bay water from EDTA, ammonium pyrollidine-dithiocarbamate,humic acid and tannic acid after treatment of the sample with ozone. Anodicscanning voltammetry was used to determine elements.

Heavy Metals, Aluminium, and Vanadium

The determination of trace metals in estuarine, marine, and other waters isthe subject of a booklet published by the Standing Committee of Analysts setup by the Department of the Environment UK [30].

The elements covered are aluminium, cadmium, chromium, cobalt, cop-per, iron, lead, manganese, nickel, vanadium, and zinc. Electrothermal atomicabsorption and anodic and cathodic scanning voltammetric methods are dis-cussed.


Newton and Van den Berg [31] applied cathodic stripping chronopotentiome-try with continuous flow to the determination of nanomolar concentrations ofnickel cobalt, copper, and uranium in estuary water.


6.11.4Emission Spectrometry

Arsenic and Antimony

Braman et al. [34] used sodium borohydride to reduce arsenic and antimonyin their trivalent and pentavalent states to the corresponding hydrides. Totalarsenic and antimony are then measured by their spectral emissions, respec-tively, at 228.8 nm and 242.5 nm. Limits of detection are 0.5 ng for antimonyand 1 ng for arsenic, copper, and silver. Oxidants interfere in this procedure.

6.11.5Hydride Generation Atomic Spectrometry

Arsenic and Antimony

It has been reported that the differential determination of arsenic [36–41] andalso antimony [42, 43] is possible by hydride generation-atomic absorptionspectrophotometry. The HGA-AS is a simple and sensitive method for the de-termination of elements which form gaseous hydrides [35, 44–47] and mg/llevels of these elements can be determined with high precision by this method.This technique has also been applied to analyses of various samples, utilis-ing automated methods [48–50] and combining various kinds of detectionmethods, such as gas chromatography [51], atomic fluorescence spectrome-try [52, 53], and inductively coupled plasma emission spectrometry [47].

Yamamoto et al. [33] applied this technique to the determination of ar-senic (III), arsenic (V), antimony (III), and antimony (V) in Hiroshima BayWater. These workers used a HGA-A spectrometric method with hydrogen-nitrogen flame using sodium borohydride solution as a reductant. For thedetermination of arsenic (III) and antimony (III) most of the elements, otherthan silver (I), copper (II), tin (II), selenium (IV), and tellurium (IV), do notinterfere in at least 30 000-fold excess with respect to arsenic (III) or anti-mony (III). This method was applied to the determination of these speciesin sea water and it was found that a sample size of only 100 ml is enough todetermine them with a precision of 1.5–2.5%. Analytical results for surfacesea water of Hiroshima Bay were 0.72 µg/l, 0.27 µg/l, and 0.22 µg/l, for arsenic(total), arsenic (III), and antimony (total), respectively, but antimony (III) wasnot detected. The effect of acidification on storage was also examined.


Heavy Metals

Beck et al. [61] used flow injection magnetic sector ICP-MS to determinecadmium, copper, nickel, zinc, and manganese in estuarine waters. The onlinepreconcentration system used Toyopearl A–T Chelate 650 H as chelating resin,and was validated for an alkaline water standard reference material (SLEW-2).


6.11.7Preconcentration Techniques

Techniques employed for the preconcentration, coastal, and estuary waters arereviewed in Table 6.2.

Table 6.2. Preconcentration techniques. Cations in estuary and coastal waters

Element Preconcentration method Analytical finish Detectionlimit, µg/l

Reference

Cr Coprecipitation with leadsulfate on lead phosphate

NAAγ -spectrometry

0.1 [54]

Sb Sb(III) and Sb(V) convertedto ammonium pyrrolidinedithiocarbamate complexesand adsorbed onto C18bonded silica, thendesorption

Graphite furnaceAAS

0.05 [55]

Cu, Cd, Pb, Zn Adsorbed ontocarboxylated polyethylene-imine-polymethylenephenyl isocyanate

ICPS – [56]

Cu, Ni, Pb, Cd Liquid–liquid extractionof ammonium pyrrolidinedithiocarbamate complexes

AAS Cu – 0.3Ni – 0.02Pb – 0.5Cd – 0.5

[57]

Fe, Mn, Cd,Zn, Cu, Ni, Pb,Co

Solvent extraction ofammonium pyrrolidinedithiocarbamate complexeswith MIBK andpreconcentration ontoChelex 100 resin

AAS 0.01–0.4 [21]

Cu, Fe Chelation–solventextraction

AAS – [25]

Miscellaneousheavy metals

Formation of ammoniumpyrrolidinedithiocarbamate complexesand solvent extraction

Flameless AAS – [33]

Fe, Mn, Cu,Ni, Cd, Pb, Zn

Adsorption onto Chelex 100resin

Flameless AAS – [58]

Cd, Cu, Fe, Pb Dithiocarbamate complexesextracted into Freon

AAS – [59]

Fe, Mn, Cu,Ni, Cd, Pb, Zn

Adsorption onto Chelex 100resin

– – [58]


References 325

6.11.8Speciation

Hasle and Abdullah [29] fractionated and speciated dissolved copper, lead,and cadmium in coastal waters from the Inner Oslofjord, Norway. They ex-amined the fractions by an operational scheme which involves ultra-filtrationfollowed by determination of labile, acid soluble and total copper, lead, andcadmium by differential pulse anodic stripping voltammetry. It was foundthat cadmium was present entirely in low molecular weight labile species; leadwas mainly in non-labile low molecular weight species, with half of the totallead probably occurring in low molecular weight organometallic compounds;copper distribution was irregular, with extensive organic and colloidal associ-ation.

Batley et al. [60] studied the speciation of the same three elements in es-tuarine and coastal waters. They evaluated the potential of a heavy metalspeciation scheme to reflect differences in metal distributions within a wa-ter mass in a study of soluble copper, lead, and cadmium speciation in watersamples from five stations in the Port Hacking Estuary (Australia) and onecoastal Pacific Ocean station. The observed metal distributions were foundto be consistent with the other measured physical and chemical properties ofthe sampled waters. In all samples, the percentages of metals associated withcolloidal matter were high, amounting to 40–60% of total copper, 45–70%of total lead, and 15–35% of total cadmium. The scheme was used to followchanges in metal speciation under different sample storage conditions. Storageat 4 ◦C in polyethylene containers was shown to prevent losses or changes inspeciation of the metals studied.

References

1. Berg BR, Abdullah MI (1977) Water Res 11:6372. Mantoura RFC, Woodward EM (1983) Estuar Coast Shelf Sci 17:2193. Loder TC, Gilbert PM (1976) Blank and salinity corrections for automated nutrient

analysis of estuarine and sea waters. University of New Hampshire Contribution UNH-59-JR101 to Technicon International Congress December 13–15

4. Amankwah D, Fasching P (1979) Anal Chem 51:11015. Epstein HS, Zander AT (1979) Anal Chem 51:9156. Gardner MJ (1987) UK Analytical Quality Control for Trace Metals in the Coastal and

Marine Environment, the Determination of Cadmium. Report PRS 1516-M, Water ResCentre, Medmenham, UK

7. Berger P, Ewald M, Lin D, Weber JH (1984) Mar Chem 14:2898. Nelson A (1985) Anal Chim Acta 169:2879. Nelson A (1985) Anal Chim Acta 169:273

10. Shuman S, Michael LC (1978) Environ Sci Technol 12:106911. Wang J, Foster N, Armalis S, Larson D, Zirino A, Olsen K (1995) Anal Chim Acta 310:22312. Knox S, Turner DR (1980) Estuar Coast Mar Sci 10:31713. Davies IM, Pirie JM (1978) Mar Pollut Bull 9:128


14. Topping G, Pirie JM, Graham WC, Shepherd RM (1975) An Examination of the HeavyMetal Levels in Muscle, Kidney, and Liver in Relation to Year, Class, Area of Samplingand Season ICE5CM, E:37

15. Topping G, Pirie JM (1972) Anal Chim Acta 62:20016. Department of the Environment and the National Water Council, Standing Commit-

tee of Analysts Manganese in Raw and Potable Waters by Spectrophotometry (usingFormaldoxime) 1977 (1978) Methods for The Examination of Waters and AssociatedMaterials, HMSO London UK

17. Willie SN, Sturgeon RE, Bermna SS (1986) Anal Chem 58:114018. Cutter GA (1978) Anal Chim Acta 98:5919. Ebdon L, Alonso JIG (1987) Analyst 112:155120. Riley JP, Skirrow G (1965) Chemical Oceanography vol 3. Academic Press, London, pp

269–40821. Sturgeon RE, Berman SS, Desauliniers JAH, Mykytiuk AP, Mclaren JW, Russell DS

(1980) Anal Chem 52:158522. Sturgeon RE, Berman SS, Desauliniers JAH, Russell DS (1979) Anal Chem 51:236423. Apte Sc, Gunn AM (1987) Anal Chim Acta 193:14724. Stein VB, Canelli B, Richards AM (1980) Int J Environ Anal Chem 8:9925. Pellenberg RE, Church TM (1978) Anal Chim Acta 97:8126. Gardner J, Yates J (1979) The determination of dissolved cadmium and lead in estuarine

water samples. CEP Consultants Ltd., Edinburgh UK, pp 427–43027. Clem RG, Hodgson AT (1978) Anal Chem 50:10228. Batley GL (1981) Anal Chim Acta 124:12129. Hasle JR, Abdullah MI (1981) Mar Chem 10:48730. Standing Committee of Analysts Department of the Environment UK (1987) Methods

for the Determination of Metals31. Newton MP, Van den Berg CMG (1987) Anal Chim Acta 199:5932. Kingston HM, Barnes IL, Brady TJ, Rains TC (1978) Anal Chem 50:206433. Yamamoto M, Urata K, Murashige K, Yamamoto Y (1981) Spectrochim Acta 36B:67134. Braman RS, Justen LL, Foreback CC (1972) Anal Chem 44:219535. Arey FK, Chamblce JW, Heckel E (1980) Int J Environ Anal Chem 7:28536. Agget J, Aspell AC (1976) Analyst 101:34137. Braman RS, Johnson DL, Foreback CC, Ammins JM, Bricher JL (1977) Anal Chem 49:62138. Andreae MO (1977) Anal Chem 49:82039. Shaikh Au, Tallman DE (1978) Anal Chim Acta 98:25140. Feldman C (1979) Anal Chem 51:66441. Nakashima S (1979) Analyst 104:17242. Yamamoto M, Urata K, Yamamoto Y (1981) Anal Lett 41:2143. Nakashima S (1980) Analyst 105:73244. Yamamoto Y, Kumamaru T, Hayashi Y, Tsujino R (1972) Anal Lett 5:41945. Yamamoto Y, Kumamaru T, Hayashi Y, Tsujino R (1972) Anal Lett 5:71746. Yamamoto Y, Kumamaru T (1976) Fresen Z Anal Chem 281:35347. Yamamoto Y, Kumamaru T (1976) Fresen Z Anal Chem 282:19348. Chan CY, Baig MWA, Pitts AE (1979) Anal Chim Acta 111:16949. Peter E, Growcock G, Strung G (1979) Anal Chim Acta 104:17750. Agemian H, Cheam V, Thomson R (1978) Anal Chim Acta 101:19351. Skogerboe RK, Bejmuk AP (1977) Anal Chim Acta 94:29752. Tsujii K, Kuga K (1978) Anal Chim Acta 97:5153. Nakahara T, Kobayashi S, Wakisaka T, Musha S (1980) Appl Spectros 34:19454. Zhang D, Smith RS (1983) Anal Chem 55:100

References 327

55. Sturgeon RE, Willie SN, Berman SS (1985) Anal Chem 57:656. Horvath Z, Barnes RM (1986) Anal Chem 58:135257. Apte SC, Gunn AM (1987) Anal Chim Acta 193:14758. Paulson AJ (1986) Anal Chem 58:18359. Danielsson L-G, Magnusson B, Westerlund S, Zhang K (1982) Anal Chim Acta 144:18360. Batley GE, Kramer KJM, Vinhas T (1996) Fresen J Anal Chem 354:39761. Beck NG, Franks RP, Bruland KW (2002) Anal Chim Acta 455:11

7 Radioactive Elements

7.1Naturally Occurring Cations

7.1.1Actinium

Lin [1] used coprecipitation with lead sulfate to separate 237actinium fromseawater. The 237actinium was purified by extraction with HDEHP and deter-mined by α-spectrometry with a Si (Au) surface barrier detector. The methodhas a sensitivity of 10–3 Ci/g of ashed sample.

7.1.2Polonium and Lead

Skwarzek and Bojanowski [2] in a study of the accumulation of 210polonium inSouthern Baltic seawater showed that the mean concentration was 0.49 mBq/dm3, of which approximately 80% was dissolved. 210Polonium concentrationsin phytoplankton and zooplankton were 21–61 and 21–451 mBq/g dry weight,respectively. Mean 210polonium concentration factors were 5000 in phyto-plankton, 18 300 in macrozooplankton, and 42 000 in mesozooplankton. Thehigher mean 210polonium concentration in mesozooplankton from the SlupskTrough compared with that in mesozooplankton from the Gdansk basin (214against 55 mBq/g dry weight) might have been due to blue-green algal bloomsin the Gdansk basin.

Various workers [3–7] have discussed the determination of polonium andlead in seawater.

Similar affinity of polonium and plutonium for marine surfaces implies thatstudies of the more easily measured polonium might be valuable in predict-ing some consequences of plutonium disposal in the oceans [8–11]. Rates atwhich plutonium and polonium deposit out of seawater onto surfaces of gi-ant brown algae and “inert” surfaces, such as glass and cellulose, suggest thatboth nuclides are associated in coastal seawater with colloidal sized specieshaving diffusivities of about 3 ×10–7 cm2/s. The parallel behaviour possibly

330 7 Radioactive Elements

represents an initial step in the incorporation of both α-radioactive heavy el-ements into marine food chains and/or their transport by the greater activityconcentrations found on marine surfaces and in seawater, about 200 times thatof plutonium.

Tsunogai and Nozaki [6] analysed Pacific Oceans surface water by con-secutive coprecipitations of polonium with calcium carbonate and bismuthoxychloride after addition of lead and bismuth carriers to acidified seawa-ter samples. After concentration, polonium was spontaneously deposited ontosilver planchets. Quantitative recoveries of polonium were assumed at the ex-traction steps and plating step. Shannon et al. [7], who analysed surface waterfrom the Atlantic Ocean near the tip of South Africa, extracted poloniumfrom acidified samples as the ammonium pyrrolidine dithiocarbamate com-plex into methyl isobutyl ketone. They also autoplated polonium onto silvercounting disks. An average efficiency of 92% was assigned to their procedureafter calibration with 210Po–210Pb tracer experiments.

Shannon [3] determined 210polonium and 210lead in seawater. These twoelements are extracted from seawater (at pH 2) with a solution of ammoniumpyrrolidine dithiocarbamate in isobutyl methyl ketone (20 ml organic phaseto 1.5 litres of sample). The two elements are back-extracted into hydrochloricacid and plated out of solution by the technique of Flynn [12] but with useof a PTFE holder in place of the Perspex one, and the α-activity deposited ismeasured. The solution from the plating-out process is stored for 2–4 months,then the plating-out and counting are repeated to measure the build-up of210polonium from 210lead decay and hence to estimate the original 210Pb ac-tivity.

Nozaki and Tsunogai [4] determined 210lead and 210polonium in seawater.The 210lead and 210polonium in a 30–50 litre sample are coprecipitated withcalcium carbonate together with lead and bismuth and are then separatedfrom calcium by precipitation as hydroxides. The precipitate is dissolved in0.5 M hydrochloric acid, and 210polonium is deposited spontaneously fromthis solution on to a silver disk and is determined by α-spectrometry. Chem-ical yields of lead and bismuth are determined in a portion of the solutionfrom which the polonium has been deposited; hydroxides of lead and othermetals are precipitated from the remainder of this solution and after a periodexceeding three months, the 210polonium produced by decay of 210lead is de-termined as before. The activity of 210lead is calculated from the activity of210polonium. The method was used to determine the vertical distribution oflead and 210polonium activities in surface layers of the Pacific Ocean.

Cowen et al. [5] showed that polonium can be electrodeposited onto carbonrods directly from acidified seawater, stripped from the rods and auto-platedonto silver counting disks with an overall recovery of tracer of 85 ± 4% for anelectrodeposition time of 16 h [13].

These workers compared two procedures for concentrating 210poloniumfrom seawater:

7.1 Naturally Occurring Cations 331

1. Coprecipitation upon partial precipitation of the natural calcium and mag-nesium with sodium hydroxide.

2. Electrodeposition of polonium directly from acidified seawater onto carbonrods.

Polonium thus concentrated was autoplated onto silver counting disks held inspinning Teflon holders.

Recoveries of 208polonium tracer in the precipitation method were 77 ± 7%(n = 8) compared with 40 ± 2% (n = 2) for the electrodeposition methodwith 16 h plating time, 64 ± 1% (n = 2) in 24 h, and 85 ± 4% (n = 2) in 48 h.

Biggin et al. [115] have described a time efficient method for the determina-tion of 210lead, 210bismuth, and 210polonium activities in seawater using liquidscintillation spectrometry.

7.1.3Radium

Orr [14] has proposed a scintillation and β–γ coincidence methods for thedetermination as measured by 220radon emanation.

Burnett and Tai Wei-Chieh [15] used α liquid scintillation to determineradium radionucleides in seawater. The method was applied in the 7–35 dpm100 kg–1-range using 1 litre samples.

Cohen and O’Nions [16] determined femtogram quantities of radium ra-dionucleides in seawater by thermal ionisation mass spectrometry.

Bettoli et al. [17] has described a shipboard system to measure the concen-trations of 222radium and 226radium in sea and coastal waters.

7.1.3.1Radium, Barium, and Radon

Perkins [18] carried out radium and radiobarium measurements in seawaterby sorption and direct multidimensional gamma-ray spectrometry. The proce-dure described includes the removal of radium and barium from water sampleson sorption beds of barium sulfate impregnated alumina (0.5–1 cm thick) anddirect counting of these beds on a multidimensional γ -ray spectrometer. Theradioisotopes can be removed at linear flow rates of sample of up to 1 m/min.

Oceanographers have developed methods to measure the 228radium con-tent of seawater, as it is a useful tracer of mixing in the ocean. These pro-cedures are based on concentrating radium from a large volume of seawater,removing all 228thorium from the sample and ageing the sample while a newgeneration of 228thorium partially equilibrates with 228radium. After storageperiods of 6–12 months, the sample is spiked with 230thorium and after ion ex-change and solvent exchange separations, the thorium isotopes are measuredin a γ -ray spectrometer system utilising a silicon surface barrier detector.


Early work was based on concentrating the radium from the seawater sampleby adding barium and coprecipitating with barium sulfate. This concentra-tion procedure has been replaced by one involving the extraction of radiumfrom seawater on acrylic fibre coated with manganese dioxide [19, 20] (Mnfibres). By use of this technique, volumes of 200–2000 litres may be sampledroutinely.

Measurements of 226radium are simpler than those for 228radium and aremore precise. These measurements are generally made by concentrating theradium from up to a few litres via barium sulfate precipitation followed by thicksource α counting or by 222radon extraction following dissolution of bariumsulfate [21].

Oceanographersusedifferent techniques formeasuring 226radiumin seawa-ter. Some workers store the sample in a 20 l glass bottle and extract successivegenerations of 222radon [22,23]. Others quantitatively extract the radium ontomanganese fibre and measure radon directly emanating from the manganesefibre [24] or in a hydrochloric acid extract from the fibres. The 222radon activ-ity is then determined by α-scintillation counting. All of these techniques givehigh levels of reproducibility and accuracy as determined by the oceanographicconsistency of the results [22, 23].

The introduction of high-resolution, high-efficiency γ -ray detectors com-posed of lithium-drifted germanium crystals has revolutionised γ -measure-ment techniques. Thus, γ -spectrometry allows the rapid measurement of rel-atively low-activity samples without complex analytical preparations. A tech-nique described by Michel et al. [25] uses Ge(Li) γ -ray detectors for the si-multaneous measurements of 228radium and 226radium in natural waters. Thismethod simplifies the analytical procedures and reduces the labour while im-proving the precision, accuracy, and detection limits.

In this method the radium isotopes are preconcentrated in the field from 100to1000 litrewater sampleontomanganese impregnatedacrylicfibre cartridges,leached from the fibre and coprecipitated with barium sulfate. This manganesefibre γ -ray technique is shown to be more accurate than the 228actinium meth-ods recommended by the Environmental Protection Agency [26] and as accu-rate but more rapid than the 228thorium ingrowth procedure.

Key et al. [27] have described improved methods for the measurementof radon and radium in seawater and marine sediments using manganesedioxide impregnated fibres. The basic method that these workers used wasthat of Broecker [28]. Seawater samples were taken in 30 litre Niskin bottles.

7.1.3.2Radium, Thorium, and Lead

226Radium, 230thorium, and 210lead in large volumes of seawater have beencollected on manganese oxhydroxide-impregnated cartridges prior to deter-mination by radiochemical methods [29].


Colley and Thomas [29] determined 226radium, 230thorium, and 210lead inlarge sample volumes of seawater. In situ pumps were used to collect particleson a 1 µm filter. The dissolved species were collected on manganese oxyhy-droxide impregnated cartridges prior to final analysis for nuclides.

Baskaran et al. [30] pumped seawater at 35 l/min and collected dissolvedspecies on cartridges prior to determining radium, thorium, and lead by γ

counting methods.

7.1.499Technetium

Chen et al. [31] preconcentrated 99technetium in seawater on an anion ex-change column to determination in amounts down to 3 mBq/m3.

Ballestra et al. [32] described a radiochemical measurement for determina-tion of 99technetium in rain, river, and seawater, which involved reduction totechnetium (IV), followed by iron hydroxide precipitation and oxidation to theheptavalent state. Technetium (VII) was extracted with xylene and electrode-posited in sodiumhydroxide solution.The radiochemical yieldwasdeterminedby gamma counting on an anticoincidence shield GM–gas flow counter. Theradiochemical yield of 50 to 150 litre water samples was 20–60%.

99Technetium has been determined in seawater by inductively coupledplasma mass spectrometry after preconcentration by coprecipitation with ironhydroxide [33].

Keith-Roach et al. [114] has described a radiochemical separation and ICPSprotocol for determining 99technetium in seawater.

7.1.5Thorium

In recent years, there has been an increasing level of interest in the use of234Th/238U disequilibrium in the marine environment to study geochemicalprocesses with short time scales (up to 100 days), particularly those associatedwith carbon cycling in the oceans [34–36] and the partitioning of pollutantsbetween the dissolved and particulate phases [37, 38]. However, the analysisof 234thorium is constrained by its short half-life and its low concentration inseawater, so appropriate analytical techniques must be rapid and sensitive andpreferably should allow shipboard analysis.

Secondary ion mass spectrometry has been used to determine low levels of230thorium and 232thorium in seawater [39].

Thermal ionisation mass spectrometry has been used to determine pg/kglevels of 230thorium and 232thorium in seawater [40].

Pates et al. [41] have described a liquid scintillation spectrometry methodfor the determination of 234thorium in seawater with 230thorium as the yield


tracer. 234Thorium is separated from the dissolved phase by a ferric hydrox-ide precipitation and is then purified using ion exchange chromatography. Thecounting source is prepared by taking the sample to dryness in a vial, redissolv-ing in acid, and mixing with a scintillation cocktail. The instrument employedhas a relatively low background (11 cpm) and the ability to separate α fromβ activity on the basis of pulse shapes. The 234thorium + 234mprotactiniumcounting efficiency is 50% over the counting window employed. The limitof detection, using the above parameters, a 20 litre sample, and a 400-mincount, is found to be 0.04 dpm/l. It was also demonstrated that less advancedinstruments, without α/β separation, can also be used effectively.

Bacon and Anderson [42] determined 230thorium and 228thorium concen-trations, in both dissolved and particulate forms, in seawater samples fromthe eastern equatorial Pacific. The results indicate that the thorium isotopesin the deep ocean are continuously exchanged between seawater and particlesurfaces. The estimated rate of exchange is fast compared with the removalrate of the particulate matter, suggesting that the particle surfaces are nearlyin equilibrium with respect to the exchange of metals with seawater.

Because of the large volumes of water that were required, an in situ samplingprocedure was used. Submersible, battery powered pumping systems [43, 44]were used to force the water first through filters (62 µm mesh Nitex followedby 1.0 µm pore-size Nucleopore) then through an adsorber cartridge packedwith Nitex netting that was coated with manganese dioxide to scavenge thedissolved thorium isotopes, and finally through a flow meter to record thevolume of water that was filtered. Natural 234thorium served as the tracer formonitoring the efficiency of the adsorbed cartridges. Standard radiochem-ical counting techniques were used [45]. On average 4% of the 234thorium,15% of the 228thorium, and 17% of the 230thorium were found in the par-ticulate form, i.e., the percentage increases with increasing radioactive half-life. However, the percentages varied considerably from sample to sampleand were found to be strongly dependent on total suspended matter concen-tration.

Traditionally, 234thorium has been analysed by gas proportional count-ing of β-particles emitted by 234mprotactinium, using sample volumes rang-ing between 20 and 100 litres, depending on detector efficiency and back-ground [38, 46, 47]. Since the analysis requires preconcentration and purifica-tion of the sample and electrodeposition onto a planchette, a yield monitor isrequired – typically 228thorium, 229thorium, or 230thorium. The samples thenrequire a minimum of two counts – one to determine the β activity and anotherto determine the α activity – and each count requires independent calibrationof detection efficiency. Modern gas proportional counting instruments are ca-pable of providing low backgrounds (–0.3 cpm) and extremely good accuracyand precision (–2%) and have been used at sea [47].

An alternative approach is γ spectrometry using HpGe γ photon detec-tors. This technique meets most of the requirements for 234thorium analysis


since no chemical manipulation of the sample is required and the detectorsare sufficiently rugged to be used at sea. However, the low absolute intensityof the 63 keV γ photon emissions (3.8%), combined with the relatively lowdetection efficiency of γ spectroscopy systems, results in large sample vol-umes (300–600 litres) being required for the analysis [48, 49]. These samplesizes can be achieved through the use of in situ pumps and manganese car-tridges [49] which scavenge thorium from seawater [50]. These systems avoidthe problems of bottle-associated sampling artefacts, such as thorium lossesto the vessel walls and particles sinking below spigots, and enable sampling ofrare large particles. However, the pumping system is relatively expensive andtime-consuming to use, restricting the number of depths that can be sampledsimultaneously. In addition, only a simple split of particulate and dissolvedfractions can be archieved, with more detailed size fractionation, such as thatrequired for the determination of colloidal 234thorium, requiring an alterna-tive method. This second point is of particular relevance with the growingrealisation that (i) colloids play a critical role in both carbon cycling and tracemetal scavenging, and (ii) 234thorium/238uranium disequilibrium is a usefultechnique for elucidating this role [51–53].

Liquid scintillation spectrometry is a technique suitable for the analysis ofboth α and β emitters, with much higher detection efficiencies than either α orγ spectrometry using semiconductor detectors or gas proportional counting.For α emitters, the liquid scintillation spectrometric detection efficiency is∼100%, while for β emitters with Emax > 156 keV (14C), detection efficiency is> 95%.

7.1.6Bromide

Foti [54] has studied the feasibility of concentrating traces of radioactive bro-mide ions by passing the seawater sample through a column of inactive AgBr(to effect isotopic exchanges). The effects of column height and of flow rate,volume and/or residence time of the seawater on the extent of exchange wereexamined; each of these variables had a significant effect.

7.1.7Phosphate

Flynn and Meeham [55] have described a solvent extraction phosphomolyb-date method using iso-amyl-alcohol for monitoring the concentration of 32-phosphorus in sea and coastal waters near nuclear generating stations.

Benitez-Nelson and Buesselet [56] have developed a new method for thecollection, purification, and measurement of natural levels of 32P and 33Pin marine particulates, and dissolved constituents of seawater. 32P and 33P


activities were measured using a ultra-low-level liquid scintillation counter.Measurement by liquid scintillation counting allows, for the first time, simul-taneous measurement of both 32P and 33P. Furthermore, 33P activities aremeasured with high efficiency (> 50%), regardless of the amount of stablephosphorus in the sample. Liquid scintillation also produces energy specific β

spectra which has enabled us to identify previously unrecognised β-emittingcontaminants in natural samples. In order to remove these contaminants,new methods of purification have been developed which utilise a series ofprecipitations and anion and cation exchange columns. Dissolved seawatersamples were extracted from large volumes of seawater (> 5000 litres) us-ing iron-impregnated polypropylene filters. On these filters, it was possibleto load between 25 and 30% Fe(OH)3 by weight, over twice that loaded onpreviously utilised materials. Using these collection, purification, and liquidscintillation counting techniques, it was possible to obtain specific 32P and 33Pactivities with less than 10% error (2a) in rainwater and 20% error (2a) inseawater.

7.2Fallout Products and Nuclear Plant Emissions

7.2.1Americium and Plutonium

Livingston and Cochran [50] collected large seawater samples by using a cable-supported electrical pumping system for subsequent determination of tho-rium, americium, and plutonium isotopes. Particles were removed by filtra-tion and actinides were collected by absorption on manganese dioxide-coatedfilters. The samples were then analysed by standard radiochemical and a spec-trometric techniques.

Schell et al. [57]havedescribeda sorption technique for samplingplutoniumand americium, from up to 4000 litres of water in 3 h. Battelle large-volumewater samples consisting of 0.3 µm Millipore filters and sorption beds of alu-minium oxide were used. Particulate, soluble, and presumed colloidal fractionsare collected and analysed separately. The technique has been used in freshand saline waters, and has proved to be reliable and comparatively simple.

7.2.2137Caesium

Dutton [58] has described a procedure for the determination of 137caesium inwater. This procedure comprises a simple one-step separation of the radio-caesium from the sample using ammonium dodecamolybdophosphate orpotassium cobaltihexacyanoferrate; 137caesium and 134caesium are measured

7.2 Fallout Products and Nuclear Plant Emissions 337

by γ -ray counting of the dried adsorbent with a Nal(Tl) crystal coupled toa γ -ray spectrometer. Levels of 137caesium activity down to about 1 pCi perlitre can be determined in seawater and lake, rain, and river waters withoutsophisticated chemical processing.

A further method for the determination of caesium isotopes in saline wa-ters [60] is based on the high selectivity of ammonium cobalt ferrocyanidefor caesium. The sample (100–500 ml) is made 1 M in hydrochloric acid and0.5 M in hydrofluoric acid, then stirred for 5–10 min with 100 mg of the fer-rocyanide. When the material has settled, it is collected on a filter (pore size0.45 µm), washed with water, drained dried under an infrared lamp, coveredwith plastic film and β-counted for 137caesium. If 131caesium is also present,the γ -spectrometric method of Yamamoto [61] must be used. Caesium can bedetermined at levels down to 10 pCi/l.

Mason [62] has described a rapid method for the separation of 137caesiumfrom a large volume of seawater. In this procedure the sample (50 litres) isadjusted to pH 1 with nitric acid, and ammonium nitrate (100 g) and cae-sium chloride (30 mg) added as carrier. A slurry is prepared of ammoniummolybdophosphate (7.5 g) and Gooch-crucible asbestos (715 g) with 0.01 Mammonium nitrate and deposited by centrifugation on a filter paper fittedin the basket of a continuous-flow centrifuge. The sample is centrifuged at600–3000 rpm and the deposit washed on the filter with 1 M ammonium ni-trate (60–70 ml) and 0.01 M nitric acid (30–40 ml). The caesium collectedon the filter is then prepared for counting by the method of Morgan andArkell [63]. With this method the caesium can be extracted in less than anhour.

Krosshaven et al. [64] used scintillation spectrometry employing germa-nium detectors to measure 137caesium and 90strontium in coastal seawaters.

Aleksan’yan [65] has discussed a method for determining 90strontium and137caesium in seawater or river water involving isolation of the radionucleides,in the presence of strontium and caesium carriers, by precipitation as thecarbonate and ferrocyanide respectively. The carbonate is dissolved in 0.5 Nhydrochloric acid and the strontium in this solution is precipitated as oxalate,the precipitation ignited and a solution of the product in 2 N hydrochloric acidis set aside for accumulation of 90yttrium; this is precipitated as hydroxide(and again as oxalate) for β-counting.

The ferrocyanide precipitate containing 137caesium is ignited at < 400 ◦C,the residue is extracted with boiling water, then evaporated and a solution ofthe residue in acetic acid is treated to precipitate caesium (as Cs3Bi2I9) forβ-counting.

Huang et al. [66] removed 137caesium from 4 litre samples of seawater byadsorption onto a filter coated with copper (II) and then determined it by with47% recovery by γ -ray spectrometry.

Riel [67] studied in situ extraction combined with γ -ray spectrometry in anunderwater probe for the determination caesium and chromium in seawater.


7.2.360Cobalt

Hiraide et al. [68] used continuous flow coprecipitation-floatation for the ra-diochemical separation of 60cobalt from seawater. The 60cobalt activity wasmeasured by liquid scintillation counting with greater than 90% yield anda detection limit of 5 fCi/l seawater.

Tseng et al. [69] determined 60cobalt in seawater by successive extrac-tions with tris(pyrrolidine dithiocarbamate) bismuth (III) and ammoniumpyrrolidine dithiocarbamate and back-extraction with bismuth (III). Filteredseawater adjusted to pH 1.0–1.5 was extracted with chloroform and 0.01 Mtris(pyrrolidine dithiocarbamate) bismuth (III) to remove certain metalliccontaminants. The aqueous residue was adjusted to pH 4.5 and re-extractedwith chloroform and 2% ammonium pyrrolidine thiocarbamate, to removecobalt. Back-extraction with bismuth (III) solution removed further trace el-ements. The organic phase was dried under infrared and counted in a ger-manium/lithium detector coupled to a 4096 channel pulse height analyser.Indicated recovery was 96%, and the analysis time excluding counting was50-min per sample.

7.2.455Iron

Testa and Staccioli [70] used Microthene-710 (microporous polyethylene) asa support material for bis-(2-ethylhexyl) hydrogen phosphate in the determi-nation of 55iron in environmental samples.

7.2.554Manganese

54Manganese has also been determined by a method [71] using coprecipitationwith ferric hydroxide. The precipitate is boiled with hydrogen peroxide and theiron is removed by extraction with isobutyl methyl ketone. Zinc is separated onan anion exchange column and the manganese is separated by oxidising it topermanganate in the presence of tetraphenylarsonium chloride and extractingthe resulting complex with chloroform prior to a spectrophotometric finish.Both 65zinc and 54manganese are counted with a 512-channel analyser witha well-type Nal(T1) crystal (7.6 × 7.6 cm). Recoveries of known amounts of65zinc and 54manganese were between 74% and 84% and between 69% and74%, respectively.

Flynn [72] has described a solvent extraction procedure for the determina-tion of 54manganese in seawater in which the sample with bismuth, cerium, andchromium carriers, is extracted with a heptane solution of bis(2-ethylhexyl)phosphateand themanganeseback-extractedwith1 Mhydrochloricacid.After


oxidation with nitric acid and potassium chlorate, manganese is determinedspectrophotometrically as permanganate ion.

7.2.6237Neptunium

May et al. [73] used neutron activation analysis to determine 237neptuniumin waste waters. The determination used the 237Np(n, γ )238Np reaction. Thedetection limit was 5 ×10–6 µg of 237neptunium, which corresponds to2.5 ×10–6 µg/kg for 200 ml seawater samples.

Holm et al. [74] used α spectrometry for the determination of 237neptuniumin seawater. The actinides are preconcentrated from a large seawater sample byhydroxide precipitation. The neptunium was isolated by ion exchange, fluorideprecipitation, and extraction with TTA. 238Neptunium or 235neptunium wasused to determine the radiochemical yield.

7.2.7Plutonium

The plutonium concentration in marine samples is principally due to environ-mental pollution caused by fallout from nuclear explosions and is generally atvery low levels [75]. Environmental samples also contain microtraces of natu-ral α emitters (uranium, thorium, and their decay products) which complicatethe plutonium determinations [76]. Methods for the determination of pluto-nium in marine samples must therefore be very sensitive and selective. Themethods reported for the chemical separation of plutonium are based on ionexchange resins [76–80] or liquid–liquid extraction with tertiary amines [81],organophosphorus compounds [82, 83], and ketones [84, 85].

Wong [77] has described a method for the radiochemical determinationof plutonium in seawater, sediments, and marine organisms. This procedurepermits routine determinations of 239plutonium activities (dpm) down to0.004 dpm per 100 litres of seawater (50 litre sample), 0.02 dpm per kg sed-iments (100 g samples), and 0.002 dpm per kg of organisms (1 kg sample).The plutonium is separated from seawater by coprecipitation with ferric hy-droxide and from dried sediments or ashed organisms by leaching with nitricacid and hydrochloric acid [86]. After further treatment and purification byion exchange, the plutonium is electro-deposited onto stainless steel disks forcounting and resolution of the activity by α-spectrometry. For 30 samples theaverage deviation was generally well within the 1SD counting error. For seawa-ter the average recovery was 52 ± 18% and for sediments and organisms it was63 ± 20%. The most serious interference is from 228thorium, which is presentin most samples and is also a decay product of the 236plutonium tracer.

Livingstone et al. [87] carried out double tracer studies to optimise con-ditions for the radiochemical separation of plutonium from large volumes ofseawater.


In this procedure 242plutonium is added to determine the overall recoveryof plutonium from the sample, and the recovery of 242plutonium at any point inthe procedure is measured by the addition of a similar amount of 236plutoniumat that point, the final recovery of 236plutonium being used to calculate the re-covery of 242plutonium at the time of the addition of 236plutonium. Experiencewith this double-tracer experiment has permitted improvement in the abilityto recover plutonium from 50 litre samples for α-spectrophotometric analysisof 239plutonium, 240plutonium, and sometimes 238plutonium.

Anderson and Fleer [88] determined the natural actinides 237actinium,228thorium, 230thorium, 232thorium, 234thorium, 231palladium, 238uraniumand 234uranium, and the α-emitting plutonium isotopes in samples of sus-pended marine particulate material and sediments. Analysis involves totaldissolution of the samples to allow equilibration of the natural isotopes withadded isotope yield monitors followed by coprecipitation of hydrolysable met-als at pH 7 with natural iron and aluminium acting as carriers to removealkali and alkaline earth metals. Final purification is by ion exchange chro-matography (Dowex AG1-X3) and solvent extraction for palladium. Over-all chemical yields generally range from 50% to 90%. The method has beensuccessfully interfaced with methods to include the determination of falloutelements of 55iron, 137caesium, 90strontium and 241americium on the samesamples.

Testa and Staccioli [70] have pointed out that Microthene-710 (a micro-porous polyethylene) as a support material for triphenylphosphine oxide incyclohexane medium has a potential application for the determination of plu-tonium in fallout samples.

Hirose and Sugimura [89] investigated the speciation of plutonium in sea-water using adsorption of plutonium (IV)-xylenol orange and plutonium-arsenazo (III) complexes on the macroreticular synthetic resin XAD-2. Xylenolorange was selective for plutonium (IV) and arsenazo (III) for total plutonium.Plutonium levels were determined by α-ray spectrometry.

Kim et al. [116] determined plutonium isotopes in seawater by an onlinesequential injection technique with sector field ICP-MS.

Delle Site et al. [90] have used extraction chromatography to determine plu-tonium in seawater sediments and marine organisms. These workers used dou-ble extraction chromatography with Microthene-210 (microporous polyethy-lene) supporting tri-octylphosphine oxide, a technique that has been usedpreviously to isolate plutonium from other biological and environmental sam-ples [91]. 236Plutonium and 242plutonium were tested as the internal stan-dards to determine the overall plutonium recovery, but 242plutonium wasgenerally preferred because 236plutonium has a shorter half-life and an α-emission (5.77 MeV) which interferes strongly with the 5.68 MeV (95%) α-line of 224radium, the daughter of 228thorium. However, the 5.42 MeV-lines of228thorium interfere with those of 238plutonium (5.50 MeV) and so a completepurification from thorium isotopes is required.


Plutoniumsourceswerecountedbyanα-spectrometerwithgoodresolution,background, and counting yield. The counting apparatus used had a resolutionof 40 keV. The mean (± sd) background value was 0.0004 ± 0.0003 cpm in the239plutonium and 240plutonium energy range and 0.0001 ± 0.0001 cpm in the238plutonium energy range. The mean (iso) counting yield, obtained with239plutonium, 240plutonium reference sources counted in the same geometry,was found to be 25.08 ± 0.72%.

To determine the overall recovery obtained by this procedure (chemicalrecovery and electrodeposition yield) a known activity of 242plutonium wasadded to the different samples; the plutonium sources were counted by α-spectrometry for 3000 minutes and the percentage overall recovery was calcu-lated from the area of 242plutonium peak. The percentage overall recovery for3-litre samples of seawater was 63 ± 10%. Owing to the very low activity of thesamples, the determination of 239plutonium, 240plutonium, and 238plutoniumin the reagents is very important in calculating the net activity of the radionu-clides.

The method proposed was checked by analysing some seawater referencesamples prepared by the IAEA Marine Radioactivity Laboratory (Monaco) forintercomparison programmes. The values reported by IAEA and the experi-mental values obtained by Delle Site et al. [90] were in good agreement.

Buesseler andHalverson[92]havedescribeda thermal ionisationmass spec-trometric technique for the determination of 239plutonium and 240plutoniumin seawater. The mass spectrometric technique was more sensitive than α

spectrometry by more than order of magnitude.

7.2.8106Ruthenium and Osmium

Kiba et al. [93] has described a method for determining this element in ma-rine sediments. The sample is heated with a mixture of potassium dichromateand condensed phosphoric acid (prepared by dehydrating phosphoric acid at300 ◦C). The ruthenium is distilled off as RuO4, collected in 6 M hydrochlo-ric acid-ethanol and determined spectrophotometrically (with thiourea) orradiometrically. Osmium is separated by prior distillation with a mixture ofcondensed phosphoric acid and Ce(SO4)2. In the separation of ruthenium–osmium mixtures recovery of each element ranged from 96.8 to 105.0%.

7.2.990Strontium

Silant’ev et al. [94] have described a procedure for the determination of90strontium in small volumes of seawater. This method is based on the de-termination of the daughter isotope 90yttrium. The sample is acidified with


hydrochloric acid, heated and, after addition of iron, interfering isotopes areseparated by double coprecipitation with ferric hydroxide. The filtrate is acid-ified with hydrochloric acid, yttrium carrier added, the solution set asidefor 14 days for ingrowth of 90yttrium, and Y(OH)3 precipitated from the hotsolution with carbon dioxide free aqueous ammonia. Then Y(OH)3 is re-precipitated from a small volume in the presence of hold-back carrier forstrontium, the precipitate dissolved in the minimum amount of nitric acid, thesolution heated and yttrium oxalate precipitated by adding precipitated oxalicacid solution. The precipitate is collected and ignited at 800–850 ◦C to Y2O3.The cooled residue is weighed to determine the chemical yield, then sealed ina polyethylene bag and the radioactivity of the saturated yttrium measured ona low-background β-spectrometer. If the short-lived nuclides 140barium and140lanthanium are thought to be present in the seawater sample, lanthanumcarrier is introduced after the first Y(OH)3 separation, and the system is freedfrom 140lanthanum by precipitation of the double sulfate of lanthanum andpotassium from a solution saturated with potassium sulfate.

Gordon and Larson [95] used photon activation analysis to determine87strontium in seawater. Samples (2 ml, acidified to pH 1.67 or 2.54 for storage)were filtered and freeze-dried. The residues, together with strontium stan-dards, were wrapped in polyethylene and aluminium foil and irradiated ina 30 MeV bremsstrahlung flux of γ -radiation. After irradiation, the sampleswere dissolved in 50 ml of acidified water and 87mstrontium was separated byprecipitation as strontium carbonate for counting (γ -ray spectrometer, Ge(Li)detector and multichannel pulse-height analyser). The standard deviation atthe 7 ppm strontium level was ±0.47.

Pinones determined 90strontium in seawater [96]. The seawater sample isfiltered and a known amount of strontium nitrate is added to the filter as a car-rier. Precipitation of the radiogenic elements, followed by addition of fumingnitric acid, separates strontium nitrate from the radiogenic elements. The90strontium is measured by the change in activity of the radiogenic daughter,90yttrium.

Chassery et al. [97] studied the 87Sr/86Br composition in marine sediments,observing excellent agreement between results obtained by ICP-MS and ther-mal ionisation mass spectrometry. Low level α-spectrometry with lithiumdrifted germanium detectors has been used to determine 90strontium in sea-water [59].

7.2.10Uranium

Spencer [98] has reviewed the determination of uranium in seawater.Bertine et al. [99] have discussed the determination of uranium in deep sea

sediments and water utilising the fission track technique. In this techniquea weighed aliquot (50–100 mg) of the powdered sample is made into a pel-


let with sufficient cellulose (as binder). The pellet is placed in a high-purityaluminium capsule and covered by polycarbonate plastic film (Lexan, 10 µmthick).

Adsorbing colloid flotation has been used to separate uranium from seawa-ter [101].

To the filtered seawater (500 ml; about 1.5 µg U) is added 0.05 M ferricchloride (3 ml), the pH is adjusted to 6.7 ± 0.1 and the uranium present as(UO2(CO3)3)4– is adsorbed on the colloidal ferric hydroxide which is floated tothe surface as a stable froth by the addition of 0.05% ethanolic sodium dodecylsulfate (2 ml) with an air-flow (about 10 ml min–1) through the mixture for5 min. The froth is removed and dissolved in 12 M hydrochloric acid-16 Mnitric acid (4:1) and the uranium is salted out with a solution of calciumnitrate containing EDTA, and determined spectrophotometrically at 555 nmby a modification of a Rhodamine B method. The average recovery of uraniumis 82%; co-adsorbed WO2–

4 and MoO2–4 do not interfere.

Adsorbing colloid flotation has also been used by Williams and Gillam [102].The fusion track method has also been used by Hashimoto [100]. In thismethod the uranium is first coprecipitated with aluminium phosphate [103],the precipitate is dissolved in dilute nitric acid and an aliquot of the solutionis transferred to a silica ampoule into which small pieces of muscovite areinserted before sealing. The uranium is then determined by measuring thedensity of fission tracks formed on the muscovite during irradiation of theampoule for 15 min at 80 ◦C in a neutron reactor. The muscovite is etched withhydrofluoric acid for1 hbefore thephotomicrography; thedensity is referred tothat obtained with standard solution of uranium. There is no interference fromthorium, and no chemical separations are required. An average concentrationof 3–40 ± 0.12 µg uranium per litre was obtained, in good agreement with thenormally accepted value.

Leung et al. [104] and Kim and Zeitlin [105] described a method for the sepa-ration and determination of uranium in seawater. Thoric hydroxide (Th(OH)4)was used as a collector. The final uranium concentration was measured via thefluorescence (at 575 nm) of its Rhodamine B complex. The detection limit wasabout 200 µg/l.

Korkisch and Koch [106, 107] determined low concentrations of uraniumin seawater by extraction and ion exchange in a solvent system containingtrioctyl phosphine oxide. Uranium is extracted from the sample solution (ad-justed to be 1 M in hydrochloric acid and to contain 0.5% of ascorbic acid)with 0.1 M trioctylphos-phine oxide in ethyl ether. The extract is treated withsufficient 2-methoxyethanol and 12 M hydrochloric acid to make the solventcomposition 2-methoxyethanol-0.1 M ethereal trioctylphosphine acid-12 Mhydrochloric acid (9:10:1); this solution is applied to a column of Dowex 1-X8 resin (Cl– form). Excess of trioctylphosphine oxide is removed by washingthe column with the same solvent mixture. Molybdenum is removed by elutionwith 2-methoxyethanol-30% aqueous hydrogen peroxide-12 M hydrochloric


acid (18:1:1); the column is washed with 6 M hydrochloric acid and uranium iseluted with molar hydrochloric acid and determined fluorometrically or spec-trophotometrically with ammonium thiocyanate. Large amounts of molybde-num should be removed by a preliminary extraction of the sample solution(made 6 M in hydrochloric acid) with ether.

Spectrophotometric analysis following extraction with Aliquot 336 has beenused todetermineuraniuminseawater [108].KimandBurnett [109]usedX-rayspectrometry to determine the uranium series nucleides including 238uranium,226radium and 210lead in marine phosphorites.

Bowie and Clayton [110] used γ -ray spectrometry to determine uranium,thorium, and potassium in sea bottom surveys.

Chen et al. [111] determined 238uranium, 234uranium and 232thorium inseawater by isotope dilution mass spectrometry. Uranium measurements weremade by using a 233uranium/236uranium double spike to correct for instru-mental fractionation. The 234uranium/238uranium ratio could be measuredroutinely to ±5% for 0.03 µg of total uranium in a 1-h data acquisition time,which is considerably shorter than α-counting. The 232thorium is measured to±20% for 0.001 µg of 232thorium.

7.2.11Miscellaneous

Spencer and Brewer [112] have reviewed the determination of radionucleidesin seawater and discuss sampling and storage methods together with tables ofradionucleides that have been determined in the oceans.

Becker and Dietze [113] used a micro mist nebuliser with ICP high resolu-tion mass spectrometry to determine sub-picogram/l levels radionucleides inseawater.

References

1. Lin B, Wu H, Cai Z (1987) Yuanzineng Kexue Jishu 21:4472. Skwarzek B, Bojanowski R (1988) Mar Biol 97:303. Shannon LV, Orren MJ (1970) Anal Chim Acta 52:1664. Nozaki Y, Tsunogai S (1973) Anal Chim Acta 64:2095. Cowen JP, Hodge VF, Folson TR (1977) Anal Chem 49:4946. Tsunogai S, Nozaki Y (1971) Geochem J 5:1657. Shannon LV, Cherry RD, Orren MJ (1970) Geochim Cosmochim Acta 34:7018. Hodge VF, Hoffman FL, Folsom TR (1974) Health Physics 27:299. Folsom TR, Hodge VR (1975) Mar Sci Commun 1:213

10. Folsom TR, Hodge VF, Gurney M (1975) Mar Sci Commun 1:3911. Goldberg ED, Koide M, Hodge VF (1976) Scripps Institution of Oceanography, La Jolla,

CA, USA12. Flynn A (1970) Anal Abstr 18:162413. Reid DF, Key RM, Schink DR (1979) Earth Planet Sci Lett 43:223

References 345

14. Orr JC (1988) J Geophys Res C:Oceans 93:826515. Burnett W, Tai W-C (1992) Anal Chem 64:169116. Cohen AS, O’Nions RK Anal Chem 63:270517. Bettoli MG, Cantelli L, Queirazza G, Roveri M, Tositti L, Tubertini O, Valcher S (1996)

Int J Environ Anal Chem 63:2918. Perkins RW (1969) Report of The Atomic Energy Commission, US, BNWL 1051 (Pt 2)

p 23–2719. Moore WS, Reid DF (1973) J Geophys Res 78:888020. Moore WS (1976) Deep-Sea Res 23:64721. US Environmental Protection Agency (1975) Radiochemical Methodology for Drink-

ing Water Regulations EPA 600/4-75-00522. Ku TL, Huh CA, Chen PS (1980) Earth Planet Sci Lett 49:29323. Chung Y (1980) Earth Planet Sci Lett 49:31924. Moore WS (1982) Estuar Coast Shelf Sci 12:71325. Michel J, Moore WS, King PT (1981) Anal Chem 53:188526. USEnvironmentalProtectionAgency (1976)National InterimPrimaryDrinkingWater

Regulations EPA-57019-79-00327. Key RM, Brewer RL, Stockwell JH, Guinasso NL, Schink RD (1979) Mar Chem 7:25128. Broecker WS (1965) An application of natural radon to problems in oceanic circula-

tions. In: Proceedings of the Symposium on Diffusion in the Oceans and Fresh Waters.Lamont Geological Observatory, New York, pp 116–145

29. Colley S, Thomson J (1994) Sci Total Environ 155:27330. Baskaran M, Murphy DJ, Santschi DPH, Orr JC, Sclink DR (1993) Deep-Sea Res 40:84931. Chen Q, Dahlgaard H, Nielson SP (1994) Anal Chim Acta 285:17732. Ballestra S, Bara G, Holm E, Lopez J (1987) J Radioanal Nucl Chem 175:5133. Uiomoshima N, Sayad M, Takashima Y (1993) Radiochim Acta 63:7334. Buesseler KO, Bacon MP, Cochran JK, Livingston HD (1992) Deep-Sea Res 39:111535. Coale KH, Bruland KW (1985) Limnol Oceanogr 30:2236. Schmidt S, Nival P, Reyss J-L, Baker M, Buat-Menard P (1992) Oceanologica Acta 15:22737. Huh C-A, Ku T-L, Luo S, Landry MR, Williams PM (1993) Earth Planet Sci Lett 116:15538. Kershaw P, Young A (1988) J Environ Radioactivity 6:139. Guo L, Santschi PH, Baskaran M, Zindler A (1995) Earth Planet Sci Lett 133:11740. Moran SR, Hoff JA, Buessler KO, Edwards RL (1995) Geophys Res Lett 22:258941. Pates JM, Cook GT, Mackenzie AB, Anderson R, Bury SJ (1996) Anal Chem 68:378342. Bacon MP, Anderson RF (1981) Trace metals in sea water. In: Wong CS et al. (eds)


43. Spencer DW, Sachs PL (1970) Mar Geol 9:11744. Krishnaswami S, Lal D, Somayajulu BLK, Weiss RF, Craig H (1976) Earth Planet Sci

Lett 32:42045. Anderson RF (1981) The marine geochemistry of thorium and protactinium. PhD

dissertation. Massachusetts Institute of Technology/Woods Hole Oceanographic In-stitution WH01-81-1

46. Coale KH, Bruland KW (1985) Limnol Oceanogr 32:18947. Buessler KO, Michaels AF, Siegal DA, Knap AH (1994) Global Biochem Cy 8:17948. Aller RC, De Master DJ (1984) Earth Planet Sci Lett 67:30849. Buessler KO, Cochran JK, Becou MP, Livingston HD, Casso SA, Hirschberg D, Hartman

MC, Flees AP (1992) Deep-Sea Res 39:110350. Livingston HD, Cochran JK (1987) J Radioanal Nucl Chem Ar 115:29951. Moran SB, Buessler KO (1993) J Mar Res 51:893


52. Niven SHE, Kepkay PE, Boraie A (1995) Deep-Sea Res Part II 42:25753. Huh CA, Prahl FG (1995) Limnol Oceanogr 40:52854. Foti SC (1971) Report Atomic Energy Comission USA AD73438455. Flynn WW, Meehan W (1973) Anal Chim Acta 63:48356. Benitez-Nelson C, Buessler KO (1998) Anal Chem 70:6457. Schell WR, Nevissi A, Huntamer D (1978) Mar Chem 6:14358. Dutton JWR (1970) Report of the Fisheries and Radiobiological Laboratory FRL 6.

Ministry of Agriculture, Fish and Food, UK59. Lewis SR, Shafrir HN (1971) Nucl Instrum Methods 93:31760. Janzer VJ (1973) Journal of Research of the UK Geological Survey 1:11361. Yamamoto O (1967) Anal Abstr 14:666962. Mason WJ (1974) Radiochem Radioanal Lett 16:23763. Morgan A, Arkell O (1963) Health Physics 9:85764. Krosshaven M, Skipperud L, Lien HN, Salbu B (1996) Health Physics 71:32665. Aleksan’yan OM (1972) Gidrokhim Mater 53:163 [Reference Zhur Khim IGGD 1972

(13) Abstract No.13G146, in Russian]66. Huang C-Y, Lee J-D, Tseng C-L, Lo J-M (1994) Anal Chim Acta 294:12167. Riel G (1995) In: King SE (ed) Proceedings of the workshop on monitoring nuclear

contamination in arctic seas. Naval Research Laboratory, Washington, DC, USA68. Hiraide M, Sakurai K, Mizuike A (1984) Anal Chem 56:285169. Tseng C-L, Hsieh YS, Yong MH (1985) J Radioanal Nucl Chem Lett 95:35970. Testa C, Stacciolo L (1972) Analyst 97:52771. Stah SM, Rao SR (1972) Currrent Science (Bombay) 41:65972. Flynn WW (1973) Anal Chim Acta 67:12973. May S, Engelmann C, Pinte G (1987) J Radioanal Nucl Chem 113:34374. Holm E, Aarron A, Ballestra S (1987) J Radioanal Nucl Chem 115:575. Comar CL (1976) Plutonium: Facts and Interferences. EPRI EA-43-SR76. Livingston HD, Mann DR, Bowen VT (1975) Analytical methods in oceanography. In:

Advances in chemistry, ser 147. ACS, Washington, DC, p 12477. Wong KM (1971) Anal Chim Acta 56:35578. Pillai KC, Smith RC, Folsom TR (1964) Nature 203:56879. Ballestra S, Holm E, Fukai R (1978) Presented at the symposium on the determina-

tion of radionuclides in environmental and biological materials. Central ElectricityGenerating Board, London, UK

80. Holm E, Fukai R (1977) Talanta 24:65981. Sakanous M, Nakamura M, Imai T (1971) Rapid methods for measuring radioactivity

in the environment. In: Proceedings of the symposium, Neuherberg. IAEA, Vienna,Austria, p 171

82. Statham C, Murray CN (1976) Report of the International Committee of the Mediter-ranean Ocean 23:163

83. Hampson BL, Tennant D (1973) Analyst 98:87384. Levine H, Lamanna A (1965) Health Physics 11:11785. Aakrog A (1975) Reference methods of marine radioactivity studies II. Technical

report services no. 169. IAEA, Vienna, Austria86. Chu A (1972) Anal Abstr 22:42787. Livingston HD, Mann DR, Bowen UJ (1972) Report of the Atomic Energy Commission

US COO-3563-12. Woods Hole Oceanographic Institute, Massachusetts, USA88. Anderson RF, Fleer AI (1982) Anal Chem 54:114289. Hirose K, Sugimura YJ (1985) J Radioanal Nucl Chem 92:36390. Delle Site A, Marchionni V, Testa C (1980) Anal Chim Acta 117:217

References 347

91. Testa C, Delle Site A (1976) J Radioanal Chem 34:12192. Buessler KO, Halverson JE (1987) J Environ Radioactiv 5:42593. Kiba T, Terada K, Suzuki K (1972) Talanta 19:45194. Silant’ev AN, Chumichev VB, Vakulovskii SM (1970) Trudy Inst Eksp Met Glav Uprav

Gidromet, Sluzhny Sov Minist SSSR 15 (2) [Ref: (1971) Zhur Khim 19GD (1) Abstr. No.1 G209]

95. Gordon CM, Larson RE (1970) Radiochem Radioanal Lett 5:36996. Pinones O (1986) Nucleotechnica 6:5597. Chassery S, Grousset FE, Lavaux G, Quetel CR (1998) Fresen J Anal Chem 360:23098. Spencer R (1968) Talanta 15:130799. Bertine KK, Chan LH, Turekian KK (1970) Geochim Cosmochim Acta 34:641

100. Hashimoto T (1971) Anal Chim Acta 56:347101. Kim YS, Zeitlin H (1971) Anal Chem 43:1390102. Williams WJ, Gillam AH (1979) Analyst 103:1239103. Smith J, Grimaldi O (1957) Bulletin of the US Geological Survey 1006:125104. Leung G, Kim YS, Zeitlin H (1972) Anal Chim Acta 60:229105. Kim YS, Zeitlin H (1972) Anal Abstr 22:4571106. Korkisch J, Koch W (1973) Mikrochim Acta 1:157107. Korkisch J (1972) Mikrochim Acta 687108. Barbano PG, Rigoli L (1978) Anal Chim Acta 96:199109. Kim KH, Burnett WC (1983) Anal Chem 55:796110. Bowie SHU, Clayton CG (1972) Translations of the Institution of Minerals and Metals

B 81:215111. Chen JH, Edwards RL, Wasserburg GJ (1986) Earth Planet Sci Lett 80:241112. Spencer W, Brewer PG (1970) Crit Rev Solid State Sci. Ocenographic Institute, Woods

Hole, Massachusetts, USA 1:409113. Becker JS, Dietze H-J (1999) J Anal Atom Spectros 14:1493114. Keith-Roach M, Stuerup S, Oughton DH, Dahlgaard H (2002) Analyst 127:70115. Biggin CD, Cook CT, Mackenzie AB, Pates JM (2002) Anal Chem 74:671116. Kim C-S, Kim C-K, Lee KJ (2002) Anal Chem 74:3824

8 Sample Preparation Prior to Analysis for Organics

Methods of identifying and measuring organic compounds have improvedgreatly in the past few decades. The development of the various kinds of in-strumental chromatographyhasmade identificationof componentsof complexorganic mixtures a matter of routine laboratory procedure, and the coupling ofthe gas chromatograph to the mass spectrometer has produced a tool of almostunbelievable power and versatility, one that will, in the long run, enable us tounravel even the most complex mixtures.

The concentration of organic materials in seawater is too low to merit directutilization of many of the modern analytical instruments; concentration bya factor of a hundred or more is necessary in many instances. Furthermore,the water and inorganic salts interfere with many of the analytical procedures.Separation of the organic components from seawater therefore accomplishestwo purposes; it removes interfering substances, and at the same time concen-trates enough organic matter to make analysis possible. It is not surprising thatconsiderable effort has been put into methods of separation and concentration.

Even after the organic compounds have been concentrated and separatedfrom seawater, the resulting mixture may be too complex for the analyticalmethod we wish to employ. If we have collected a large enough sample of or-ganic material, we may then resort to any of a number of fractionation schemes,based perhaps on functional groups or molecular weights, in order to simplifythe mixtures to the point where analysis could be relatively straightforward.Alternatively, if we wish to measure only one compound or class of compounds,we may try to design a concentration method which will be specific for thecompounds of interest, thus achieving concentration and fractionation in onestep. Both of these approaches have been followed with some success.

There is an extensive literature on concentration, separation, and fraction-ation methods. References [1–5] give general reviews.

The simplest approach to the collection and subdivision of organic ma-terials in seawater is to use some physical or chemical means of removingone fraction from solution or suspension. The techniques vary, from simplefiltration to collect particulate matter, to chemical methods, such as solventextraction and coprecipitation. With each of these methods, the analyst mustknow the efficiency of collection and exactly which fraction is being collected.Very often the fraction is defined by the method of collection; two methods

350 8 Sample Preparation Prior to Analysis for Organics

that purport to collect the same fraction may in fact be sampling very differ-ent universes. Comparisons between the results of different investigators areusually difficult because we do not know whether differences in results stemfrom real differences in the areas or times of collection or from differences inthe methods used. Intercalibration of analytical methods is generally agreed tobe a good thing, once a field has become sufficiently stabilized so that certainmethods are more or less recognized as “standard” methods. It is now moreoften being realised that the intercalation should start not with the analyticalmethods, but with the sampling methods, if true comparisons are to be made.

8.1Soluble Components of Seawater

Various techniques for the concentration of organics in the soluble fraction ofseawater prior to analysis are discussed below.

8.1.1Reverse Osmosis

This technique has been applied to the concentration of organochlorine andorganophosphorus insecticide [7, 8] and various ethers, glycols amines, ni-triles, hydrocarbons, and chlorinated hydrocarbons. Although this work wasconcerned with drinking water, it is a useful technique which may have applica-tion in seawater analysis. Cellulose acetate [9], ethyl cellulose acetate [6], andcrosslinked polyethyleneinine [8] have been used as semi-permeable mem-branes.

8.1.2Freeze Drying

This is another technique which has applications in seawater analysis. Approx-imately 100% recovery of glucose and lindane at the 0.1 and 0.15 mg/l levelhave been obtained from water by this technique.

Pocklington [10]has separatedaminoacids in seawaterusing this technique.The first step in his concentration of free amino acids from seawater was thefreeze drying of the seawater sample. To reduce interferences in the later stepsof the procedure, the sea salts were packed into a chromatographic columnand washed with diethyl ether to remove non-polar compounds. The diethylether extract, particularly from surface water samples, quite often containedcoloured materials as well as other organics. If a series of solvents of graduatedpolarity were passed through a sea salt column, a fractionation by polarityshould be obtained. With the proper choice of solvents, a form of gradientelution could be devised which would result in a continuous, rather thana batch fractionation.

8.1 Soluble Components of Seawater 351

8.1.3Freezing-Out Methods

Slow freezing, with constant stirring, results in a concentration of organicmaterials in the solution remaining. The technique is most effective in waterof low salinity. It has been applied to lake water with some success but marineapplications do not seem to have been developed [9, 11–13].

8.1.4Froth Flotation

In its passage through a water column, a bubble acts as an interface betweenthe liquid and vapour phases, and as such collects surface-active dissolvedmaterials as well as colloidal micelles on its surface. Thus in a well-aeratedlayer of water, the upper levels will become progressively enriched in-surface-active materials. In the open ocean, an equilibrium undoubtedly exists betweenthe materials carried downward by bubble injection from breaking waves andthose carried upward by rising bubbles. In the laboratory, however, this effectwill enrich the surface layer with organic materials.

When the concentrations of surface-active materials are high, the injectionof bubbles into the solution from well below the surface may result in theformation of a foam at the surface. The foam can be as much as 200 timesas concentrated in organic material as the body of the solution [14]. Naturalfoams of this kind can often be seen along beaches during periods of strongwinds and violent wave action. Even when the surface-active materials are notpresent at levels high enough for foam formation, a considerable enrichmentof organic material can be found in the upper part of the water column [15,16].

8.1.5Solvent Extraction

Solvent extraction is another attractive method for concentrating a particularfraction of the dissolved organic matter, the fraction concentrated being de-termined by the choice of solvent. The most obvious limitation on the methodis that set by solvent choice; the solvent should have only limited solubilityin water, which limits the materials removed to the less polar compounds.Another limitation is that set by contamination. The solvent used must be pu-rified carefully, since the amounts of the various organic compounds collectedfrom seawater will be about as large as the trace impurities in the solvents.Because of the relatively large amount of time and equipment required for theprocessing of each sample, these methods must be used to characterize theorganic compounds at a few selected stations and depths, rather than in masssurveys. Once the separation into the organic solvent has been accomplished,any of a number of techniques of fractionation and analysis can be applied.


The first step in the solvent extraction is the actual sampling of the wa-ter column. All of the problems associated with sampling can occur in thisstep, and may be aggravated by the large volumes of seawater customarilyemployed. The cleverest approach to this problem is to avoid sampling inthe normal manner altogether. Ahnoff and Josefsson [17] have described anin situ apparatus for solvent extraction. This apparatus is buoyed at the sam-pling depth, anywhere between the surface and 50 m, and water is pumpedthrough a series of extraction chambers. The capacity of the unit is 50 l per48 h. The use of an in situ pumping system on the far side of the extractionchambers eliminates the pump and hose contamination, as well as much ofthe contamination coming from the passage through the surface film. Sincethe apparatus is battery powered, the unit may be suspended from a free-floating surface buoy; no ship time is required, except for placement andrecovery of the samplers. Thus while each sample may require up to 48 hto collect, a number of depths and areas may be sampled in the same timeperiod.

Ahnoff and Josefsson [18] built a solvent extraction apparatus for river workwhich was later modified into their in situ extractor [17]. The unit as describedin the earlier work could easily be adapted for seawater analysis. A unit basedon a Teflon helix liquid–liquid extractor, some 332 feet (101.5 metres) in length,was constructed by Wu and Suffet [19]. The extractor was optimized for theremoval of organophosphorus compounds, specifically pesticides, with an ef-ficiency of around 80%. For some compounds, these continuous extractionmethods should be the methods of choice and should be explored.

The conventional approach to solvent extraction is the batch method. Earlywork with this method was hampered by the low concentration of the com-pounds present and the relative insensitivity of the methods of characteriza-tion. Thus lipids and hydrocarbons have been separated from seawater by ex-traction with petroleum ether and ethyl acetate. The fractionation techniquesinclude column and thin-layer chromatography with final characterisationby thin-layer chromatography, infrared, and ultra-violet spectroscopy and gaschromatography. Of these techniques, only gas chromatography is really usefulat levels of organic matter present in seawater. With techniques available to-day such as glass capillary gas chromatography and mass spectrometry, muchmore information could be extracted from such samples [20].

This type of separation and fractionation has been proposed by Copin andBarbier [21].

The Oil Companies International Study Group for Conservation of Clear AirandWater-Europe (Concawe) [22]havemadeadetailed studyof theapplicationof solvent extraction to the determination of organics in water.

Solvent extraction has proved to be most useful when applied to the concen-tration of particular compounds for which there exists an analytical methodof great sensitivity. The major application of the method has been for thedetermination of hydrocarbons in seawater.

8.1 Soluble Components of Seawater 353

In general, solvent extraction is an excellent method for the concentrationand determination of specific compounds, chiefly non-polar, in seawater [23].Special precautions must be taken to prevent contamination from trace ma-terials in the solvents used. In situ methods offer many advantages, not theleast being the elimination of lengthy processing in the shipboard laboratory.When coupled to modern separation and detection systems, the methods mayoffer us the simplest and most direct approach to the measurement of certainclasses of compounds. In most cases, we have little or no estimate of the ef-ficiency of the solvent extraction techniques. Because of the great variety ofcompounds present in any one sample, a true efficiency of extraction may beimpossible to obtain. Working efficiencies, using model compounds, may bethe only approach in trying to make the analysis truly quantitative.

During the investigation of pollution in coastal seawaters, Werner andWaldichuk [24] pointed out the need for concentrating and isolating traceamounts of certain substances with a continuous solvent extractor. They con-structed a modified Scheibel apparatus by changing the organic solvent cyclesystem.

8.1.6Coprecipitation Techniques

Another method of segregating and removing a portion of the dissolved or-ganic matter includes incorporation into a solid phase, which can then beremoved by filtration or centrifugation. This incorporation can result eitherfrom coprecipitation with a solid phase formed in a reaction in the solution or,as discussed in the next section, from adsorption of the organic material ontoa pre-existing solid phase.

The adsorption of organic matter on any surface presented to seawaterhas been well documented. Neihof and Loeb [25] have demonstrated thisadsorbance by following the change in surface charge of newly immersedsurfaces. There has even been an attempt to use this phenomenon as a means ofmeasuring dissolved organic carbon. Chave [26] found an association betweencalcite and dissolved organic materials in seawater, and Meyers and Quinn [27]tried to use the effect as a method for the collection of fatty acids. As a collectiontechnique adsorption on calcite has several advantages. The pH of the sampleis not greatly altered by the addition of small amounts of calcite; the precipitateis dense and should settle quickly; and after filtration the inorganic supportcan be removed by acidification. Unfortunately, the recovery of added fattyacids was inefficient; of the order of 18%. Meyers and Quinn [28] achieveda somewhat greater efficiency of collection of fatty acids with clays, but theinsolubility of the clays nullified one of the advantages of this concentrationtechnique.

The precipitant most commonly used for the collection of organics has beenwhat is loosely called “ferric hydroxide”. It is formed by the in situ formation


of hydrated ferric oxides, usually by the addition of ferrous iron, followed bypotassium hydroxide. The technique was first used for the precipitation oforganic matter from an aged algal culture [29]. They recovered 79–95% of the14C-labelled material from such cultures. Williams and Zirino [30] measuringefficiencies of removal of dissolved organic carbon, found that such scavengingcollected between 38% and 43% of the organic carbon measurable by wetoxidation with persulfate. Chapman and Rae [31] examined the effect of thisprecipitation on specific compounds. They found coprecipitation to be morecomplete with copper hydroxides, but still far from satisfactory. Only certaincompounds were removed effectively by this treatment and the efficiency ofremoval varied with the water type and the organic compound involved.

A limited amount of work has been carried out using zirconium phosphates,compounds with well-defined coagulation and adsorption properties. Theefficiency of coprecipitation was about 70% for free amino acids and albumin.

These methods may prove useful in the qualitative analysis of organic com-pounds, once the selectivities of the precipitants are understood. The metal-lic oxides suffer from the disadvantage of producing a precipitate which isdifficult to filter, while calcite and zirconium phosphates produce relativelywell-mannered precipitates. Even when the efficiencies of collection of vari-ous model compounds in seawater is known, the immense variety of organiccompounds in seawater will keep this technique largely qualitative.

8.1.7Adsorption Techniques

Coprecipitation techniques, as discussed in the previous section, are basicallybatch processes operating on a limited amount of organic matter. When reallylarge concentrations of organic matter are required, it is much more efficient,as discussed below, to hold the adsorptive material in a column, through whichthe seawater is passed. After the proper amount of water has passed throughthe column, the organic matter can be desorbed with the proper choice ofsolvents. If the proper sequence of solvents is employed, a rough fractionationof the adsorbed material can be also result.

One of the earliest choices of adsorbent for seawater organics was activatedcharcoal [29, 32]. The technique has been refined for use in both fresh andseawater by a number of workers [33–36]. A major problem in this techniquehas been the unknown efficiencies of collection and desorption. Jeffrey [37]using 14C-labelled material, found that 80% of the organic material in theseawater was adsorbed by the charcoal. Of that 80% approximately 8% againwas desorbed by the solvents used. The overall efficiency of the method ofcollection is one of the few that permits the accumulation of gram amounts oforganic materials from seawater and therefore also permits the application ofmany of the standard techniques of organic analysis. However, we do not knowhow the distribution of compounds is changed in the process of adsorption

8.2 Volatile Compounds of Seawater 355

and desorption; certain classes of compounds are probably entirely removedfrom the mixture, while others may be retained only in part. When so activea surface is used, there is also a possibility that chemical changes may takeplace during adsorption. The mixture as released from the charcoal may beconsiderably different from that original present in seawater. This collectionmethod, although attractive for its simplicity and speed, is limited to qualitativeresults.

Macroreticular resinshavealsobeenused for thecollectionof traceorganics.An excellent early review of the properties of the various XAD resins,

along with comparisons with EXP-500 and activated carbon, can be foundin Gustafson and Paleos [38].

Riley and Taylor [39] have studied the uptake of about 30 organics from sea-water onto the resin at pH 2–9. At the 2–5 µg/l level none of the carbohydrates,amino acids, proteins or phenols investigated were adsorbed in any detectableamounts. Various carboxylic acids, surfactants, insecticides, dyestuffs, and es-pecially humic acids are adsorbed. The humic acids retained on the XAD-1resin were fractionated by elution with water at pH 7, M aqueous ammonia,and 0.2 M potassium hydroxide.

Osterroht [40,41] studied the retention of non-polar organics from seawateronto macroreticular resins.

Each of the XAD resins has slightly different properties and should col-lect a slightly different organic fraction from seawater. The major differencesbetween the resins is in the degree of their polarity.

8.2Volatile Compounds of Seawater

8.2.1Gas Stripping

Whilst much of the literature on this subject is concerned with non-salinewater samples, it is believed that many of these procedures will also worksatisfactorily with seawater; indeed, the presence of salts in the sample mayassist in the removal of volatiles.

In the collection of organic materials, separation of the more volatile mate-rials can be achieved by transferring them into the vapour phase for collectionand concentration. The usual method for effecting such a transfer is to bubblesome inert gas through the liquid. The effluent gas is then passed through anadsorbent to collect the organic materials [42–47]. The desorbed material isthen analysed by gas chromatography or by linked gas chromatography-massspectroscopy.

Material removed from the water by stripping in this manner can also beconcentrated by trapping in a loop immersed in a cooling bath. The usualcooling baths are liquid nitrogen or solid carbon dioxide with or without an


organic solvent. The major disadvantage of the liquid nitrogen bath, alongwith the cost and the limited availability, particularly aboard ship, is that itcondenses carbon dioxide and water, along with the organic materials actuallydesired. These trapping techniques have been used by many workers and areusually described in conjunction with a gas chromatographic determination ofsome fraction of the organic materials. An example of this kind of separationis the work of Novak et al. [48]. They improved the yield of earth organics bysalting out the volatiles with sodium sulfate.

While the gases used in stripping are usually air, nitrogen, or helium,electrolytically evolved hydrogen has been used as a collector for hydro-carbons [49]. In this technique, the gas is not passed through a column ofadsorbent, but instead collects in the headspace of the container. Since the vol-ume of seawater and of hydrogen are known, the hydrocarbon concentrationin the headspace can be used to calculate the partition coefficients and theconcentration of hydrocarbon in the seawater.

As in most of the separation methods, the stripping and collection tech-niques have not been investigated for their recovery efficiences except in iso-lated instances. Kuo et al. [50] have found recovery efficiences for volatile polarorganics ranging from 9.5% for ethanol to 88.8% for acetone. Most of themodel compounds displayed recovery efficiencies in the 60–80% region underthe conditions of their experiment.

The Oil Companies International Study Group for Conservation of Clean Airand Water-Europe (Concawe) has made a detailed study [22] of the applicationofgas stripping to thedeterminationofhydrocarbons inamountsdowntopartsper billion in water. In this procedure the water sample is purged with nitrogenand helium and the volatiles trapped on a solid adsorbent such as carbon. Theorganics are then released from the carbon by heating and purging directly intoa gas chromatograph linked to a mass spectrometer. For chlorine and brominecontaining impurities a halide selective detector such as the Hall electrolyticconductivity detector is used on the gas chromatograph. Alternatively an alkaliflame ionization detector or an electron capture detector could be used.

Colenut and Thorburn [51,52] have also described the procedure using gasstripping of the aqueous sample followed by adsorption onto active carbonfrom which surface they are taken up in an organic solvent for gas chromato-graphic analysis. They optimized conditions for the determination of parts perbillion of pesticides and polychlorinated biphenyls.

The closed loop gas stripping system has been discussed by various work-ers [36, 53–56]. In this technique organic compounds are removed from waterby purging with a gas saturated with water vapour. Volatile and semi-volatilecompounds will partition out into the headspace and are swept to an activatedcharcoal trap. The charcoal will retain organics while allowing the purge gasto pass through. The purge gas is then returned to repurge the sample viaa pump. At the end of the purge time, typically two hours, the trap is removedand fitted with a glass collection vial. Organic compounds are extracted from


the charcoal with a small volume of a suitable solvent such as carbon disul-phide or dichloromethane, which is then collected and injected into a capillarygas chromatograph or a capillary gas chromatograph coupled with a massspectrometer.

Grob and Zurcher [36,53–55] have carried out very detailed and systematicstudies of the closed loop gas stripping procedure and applied it to the de-termination of parts per billion of 1-chloroalkanes in water. Westerdorf [56]applied the technique to chlorinated organics and aromatic and aliphatic hy-drocarbons.

Waggott [57] reported that a factor of major concern in adapting the tech-nique to more polluted samples is the capacity of the carbon filter, whichusually contains only 1.5–2 mg carbon. He showed that the absolute capacityof such a filter for a homologous series of 1-chloro-n-alkanes was 6 µg for com-plete recovery. Maximum recovery was dependent on carbon number, being ata maximum between C8 and C12 for the 1-chloro-n-alkane series. It is impor-tant, therefore, to balance the amount of sample stripped with the capacity ofthe carbon filter to obtain better than 90% recoveries.

8.2.2Headspace Analysis

Corwen [58] used this method for the analysis of ketones and aldehydes inseawater. Halocarbons were similarly separated from environmental samplesby Kaiser and Oliver [59]. There have been many other applications of thetechnique [60–69]. The major advantage of the headspace method is simplicityin handling the materials. At most, only one chemical, the salt used in thesalting-out procedure, needs to be added and in most cases the headspace gascan be injected directly into a gas chromatograph or carbon analyser. On theother hand the concentration of organic materials present is limited by thevolume of seawater in the sample bottle. This is very much a batch process.

Equilibration between the headspace gas and the solution can take a con-siderable time. This is not a problem when the salting-out material is addedat sea and the samples are then brought into the laboratory for analysis sometime later. When the salting-out is done in the laboratory, equilibration canbe hastened by recirculating the headspace gas through the solution. A sys-tem could be devised which would permit the accumulation of volatiles froma large volume of water into a relatively small headspace, perhaps by recirculat-ing both water and headspace gas through a bubbling and collection chamber,but much of the simplicity and freedom from possible contamination wouldbe lost in the process.

Volatile organic materials can also be removed from solution by distillation,either at normal or at elevated pressures. While the amounts to be collected inthis fashion are small, if headspace samples are taken at elevated temperaturesand pressures, trace quantities of organics can be detected [70]. It should be


emphasized, however, that whenever extreme conditions are employed to freean organic fraction, that fraction is defined by the conditions of the separationandcannotprofitablybe comparedwith fractionsdefinedbydifferent samplingconditions. The use of elevated temperatures and pressures may also alterthe compounds separated, limiting the amount of information that can beextracted from the analyses.

Friant and Suffet [66] have investigated in detail the interactive effects oftemperature, salt concentration, and pH on headspace analysis for isolatingvolatile trace organics in aqueous samples. Optimal conditions were derivedfrom a statistical evaluation of the effect of parameter variation on the parti-tion coefficient. These were a pH of 7.1, a sample temperature of 50 ◦C, anda salt concentration equivalent to 3.35 M sodium sulfate. Dowty et al. [63]passed the headspace purge gas through a column of Tenax GC (poly (p-2,6-diphenyl-phenylene) oxide) adsorbent to trap the organics. The organics arethen released from the Tenax GC and swept into a gas chromatograph for anal-ysis in the parts per billion range. Bellar and Lichtenberg [65] also used theprinciple of adsorption of the organic components of the purge gas on a solidadsorbent material.

Chian et al. [69] point out that the Bellar and Lichtenberg [65] procedure ofgas stripping followed by adsorption onto a suitable medium and subsequentthermal desorption onto a gas chromatograph-mass spectrometer is not verysuccessful for trace determinations of volatile polar organic compounds suchas the low molecular weight alcohols, ketones, and aldehydes. To achieve theirrequired sensitivity of parts per billion, Chian et al. [69] carried out a simpledistillation of several hundred ml of sample to produce a few ml of distillate.This achieved a concentration factor of between 10 and 100. The headspace gasinjection-gas chromatographic method was then applied to the concentrateobtained by distillation.

8.2.3Fractionation

Once a sample of dissolved organic matter has been isolated, it is still seldomin a form that permits simple analysis. In most cases, there are far too manycompounds present and some form of fractionation must take place to removeinterferences and simplify analytical procedures.

One could devise many different bases for the fractionation of organicmaterials, functional groups, degree of saturation, presence or absence ofaromatic groups, and degree of polarity have all been used. The approach mostoften used is a fractionation by size. At the upper end of the size range we aredealing with particles consisting of many discrete molecules. Fractionation isaccomplished by differential filtration, using filters and screens of decreasingpore size. A good example of the results of such fractionation is found inSheldon [71].


Particles of smaller sizes, from the colloidal to the macromolecular, areseparated by membrane filters. The most familiar of these is the Amicon Diflofilter, although several other companies now manufacture similar products.Separations in the same size range can also be achieved with hollow poly-meric fibres. At the upper end of their size range, these filters can be usedto separate different size classes of material normally considered as colloidal.At the smaller end, the separation is made on the basis of molecular size.The results are presented in terms of molecular weight, but the molecularweight calibration is done with spherical molecules. The results are thereforegiven as equivalent, spheres, rather than as true molecular weights. The tech-niques have been applied to coastal seawater. Ultrafiltration as a fractionationmethod gave recoveries of 80–100% when the carbon present in each fractionis summed.

Ultrafiltration techniques employing membrane filters and those using hol-low fibres both worked well for the concentration and desalting of humic andfulvic acids, but the high priming volume needed for the hollow fibre apparatusrestricts it to large volume applications. This is not likely to be a problem inmarine work, where large volumes are required because of the low concentra-tions of organic materials. Both membranes and fibres retained material wellbelow the expected molecular weight cut-off.

These techniques are now coming into use in marine organic chemistry.The apparatus is now available for processing large quantities of seawater,at pilot plant levels, to yield gram quantities of dissolved organic materialsin specified molecular size classes. This should be one of the most fruitfulmethods for accumulation, separation, and rough fractionation of dissolvedorganic materials.

Separation into molecular size classes by ultra-filtration is necessarily dis-continuous; the fractions resulting are composed of mixtures of compoundswithin a given band of molecular sizes. This kind of fractionation is best carriedout by some form of column chromatography. If molecular size is to be the cri-terion for separation, then materials such as Sephadex can be used as columnpacking. Sephadex separates compounds by exclusion, holding the smallermolecules within the particles and rejecting those that will not fit within thepores of the resin. Thus with a Sephadex column, the large components comeoff the column first. The system is not perfect; some charged compounds, suchas phenols, can be bound irreversibly to some of the resins. The procedurehas been used in the analysis of natural waters [72, 73] but it has not beendeveloped to its full usefulness.

XADresinshavebeenused tocollect andconcentrateorganicmaterials fromseawater. They can also be used as packings for fractionation by column chro-matography. While they have been used in simple gravity flow column chro-matography, high pressure liquid chromatography has also been used [74–78].

Reversed phase chromatography is a variant of high-performance liquidchromatography where a non-polar organic phase is immobilised and a polar


solvent is used as eluent. This variant may also be applied when non-polarcompounds are to be sorbed from a polar solvent. This very situation is en-countered in the attempt to accumulate non-polar organic substances fromseawater by liquid–solid adsorption.

In applying the principle of reversed-phase chromatography to the accumu-lation of dissolved organic material from water, Ahling and Jensen [79] useda mixture of n-undecane and Carbowax 4000 monostearate on Chromosorb Was the collecting medium. Uthe and Reinke [80] tested porous polyurethanecoated with liquid phases such as SE3, DC 200, QF-1, DEGS, OV-25, OV-225for the same purposes. In each case the coating is achieved easily and maybe modified to the desired adsorption properties. However, the coating is notchemically bonded to the support and may thus be removed together with thesorbate. Aue et al. [81] finally demonstrated the potential of support-bondedpolysiloxanes for a simple, fast, and sensitive analysis of organochlorine com-pounds in natural aqueous systems.

Derenbach et al. [82] tested a technique for the accumulation of certainfractions of dissolved organic material from seawater; and subsequently forthe fractionated desorption of the collected material.

The handling of water extracts and possible sources of contamination wouldthus be reduced to a minimum. Furthermore, fractionated desorption of theaccumulated material under mild conditions should result in less complexmixtures with little risk of denaturation.

Theseworkers investigated the suitability ofnumerous supportmaterials foruse in reversed-phase high-performance liquid chromatography for the recov-ery of non-polar organic compounds from seawater. Porous glass treated withtrichloro-n-octadecyl silane was found to permit at least a semi-quantitativerecovery of test compounds. This silanised glass support was found to be easyto keep free from contamination and in addition, had a relatively high adsorp-tion capacity, permitting fractional desorption of the test compounds. Resultsobtained with this column were compared with those obtained using AmberliteXAD-2. Derenbach et al. [82] give full details of the preparation of this supportmaterial. They used 14C labelled spike compounds, 1–14C n-hexadecane anddi(2-ethylhexyl) (carboxyl–14C)phthalate) and also non-labelled compounds(nC6-nC24 alkanes, diethyl, di-isobutyl, di-n-butyl, butylbenzyl dicyclohexyl,bisethylhexyl phthalic acid esters. p, p′-DDE DDMU Dieldrin pesticides) inrecovery experiments.

Recoveries between 103% (dialkyl phthalates) and 71% (n-alkanes, dieldrin,eldrin) were reported when the technique was applied to 25 litre samples.

A disadvantage of the reversed-phase HPLC technique or that of Derenbachet al. [82] is that it is only semi-quantitative. Against this is the fact that thetechnique using silanised porous glass is easy to keep free of contamination,has a comparatively high adsorption capacity, and permits the fractionateddesorption of accumulated material.

8.3 Chemical Pretreatment of Organics 361

Hayase et al. [83] applied reversed phase liquid chromatography with dou-ble detectors (fluorescence and absorption) to the determination of dissolvedorganic matter in estuarine seawater. The dissolved organic matter was ex-tracted into chloroform at pH 3 or 8. The hydrophilic–hydrophobic balanceand aromatic character of the seawater-dissolved organic matter was repre-sented on the chromatograms. The results indicated that reversed phase liquidchromatography with double detectors was an effective technique in the char-acterization of dissolved organic matter in seawater.

8.3Chemical Pretreatment of Organics

In some cases, the forms in which the organic materials are present are noteasily separated or identified. Some chemical pretreatment is necessary be-fore the analytical procedures can be used. These pretreatments may includethe preparation of volatile derivatives for use in gas chromatography, or thesplitting of a large and otherwise intractable molecule, as in the hydrolysis ofproteins to their constituent amino acids.

For many of the organic materials in seawater, some form of chemicalpretreatment is necessary before analysis is possible. The obvious cases are thehydrolysis of polysaccharides and proteins before the analysis for monomericconstituents, and the formation of volatile derivatives to permit analysis by gaschromatography. These methods will be discussed further in Chap. 9.

More general reactions exist which are useful in determining the skeletalstructure of unknown organic compounds. One such reaction is a treatmentwith hydrogen, resulting in the replacement of halogen, oxygen, sulfur, andnitrogen, and the saturation of double bonds [84]. The system is attached on-line in a gas chromatograph, the hydrogen being supplied by the carrier gas,the conversion taking place in a catalyst-containing tube. The resultant chro-matographic pattern is greatly simplified, making identification of chemicalstructures somewhat easier.

Carbohydrazide can be used in a similar manner as a reducing agent, con-verting azo and nitro compounds to the corresponding amines for bettervolatilization. A considerable amount of literature exists on methods of thiskind [85]; it has scarcely been used by marine chemists, perhaps because theyare still concerned with concentration and isolation of compounds in seawater.

One form of chemical pretreatment that has been used more extensively isthe oxidation of dissolved organic matter to carbon dioxide with high intensityultraviolet light. This is the basis for at least one method of measurement ofdissolved organic carbon; it has also been used for the measurement of stablecarbon isotope ratios in dissolved and particulate organic matter [86, 87].This measurement rests on the assumption that oxidation of all organics inseawater by ultraviolet light goes to completion, an assumption that is by nomeans proven.


References

1. Koyama T (1963) Journal of the Oceanographic Society of Japan 20:5632. Andelmen JB, Caruso SC (1971) Concentration and separation techniques. In: Ciaccio

LL (ed) Water and water pollution handbook, vol 2. Marcel Dekker, New York, pp483–591

3. Webb RG (1975) Isolating organic water pollutants, XAD resins, urethane foams, solventextraction. US National Technical Information Service, PB report no. 245647

4. Whitby FJ (1975) Proceedings of the Analytical Division of the Chemical Society 12:1105. Fritz JS (1977) Accounts of Chemical Research 10:676. Klein E, Eichelberger J, Ever C, Smith J (1975) Water Res 9:8077. Chian ESK, Bruce WN, Fang HHP (1975) Environ Sci Tech 9:528. Deinzer H, Melton R, Mitchell D (1975) Water Res 9:7999. Kammerer Jr PA, Lee GF (1969) Environmental Science and Technology 3:276

10. Pocklington R (1970) A new method for the determination of amino acids in sea waterand an investigation of the dissolved free amino acids of North Atlantic Ocean waters.PhD Dissertation. Dalhousie University, pp 1–102

11. Shapiro J (1961) Science 133:206312. Shapiro J (1967) Anal Chem 39:28013. Habermann HM (1963) Science 140:29214. Wallace Jr GT, Wilson DF (1969) Foam separations as a tool in chemical oceanography.

Report no. 6958. US Naval Research Laboratory, p 115. Dorman DC, Lemlich R (1965) Nature 207:14516. Karger BL, DeVivo DG (1968) Separation Science 3:39317. Ahnoff M, Josefsson B (1976) Anal Chem 48:126818. Ahnoff M, Josefsson B (1974) Anal Chem 46:65819. Wu C, Suffet IH (1977) Anal Chem 49:23120. Fielding W, Gibson TM, James HA, McLoughlin K, Steep CP (1981) Organic micropollu-

tants in drinking water. Technical report TR 159. Water Research Centre, Medmenham,UK

21. Copin G, Barbier M (1971) Oceanography 23:45522. The Oil Companies International Study Group for Conservation of Clean Air and

Water, Europe (1982) Analysis of trace substances in aqueous effluents from petroleumrefineries. Concawe report no 6/82

23. Gomella C, Belle JP, Auvray J (1976) J Techniques et Sciences Municipales 71:43924. Werner AE, Waldichuk M (1962) Anal Chem 34:167425. Neihof R, Loeb G (1974) J Mar Res 32:526. Chave KE (1965) Science 148:172327. Meyers PA, Quinn JG (1971) Limnol Oceanog 16:99228. Meyers PA, Quinn JG (1973) Cosmochim Acta 37:174529. Jeffrey LM, Hood DW (1958) J Mar Res 17:24730. Williams PM, Zirino A (1964) Nature 204:46231. Chapman G, Rae AC (1967) Nature 214:62732. Wangersky PJ (1952) Science 115:68533. Vaccaro RF (1971) Environ Sci Technol 5:13434. Grob K (1973) J Chromatogr 84:25535. Kerr RA, Quinn JG (1975) Deep Sea Res 22:10736. Grob K, Zuercher F (1976) J Chromatogr 117:28537. Jeffrey LM (1969) Development of a method for isolating gram quantities of dissolved

organic matter from seawater and some chemical and isotopic characteristics of the

References 363

isolated material. PhD Dissertation. Texas A&M University, College Station, Texas USAp 152

38. Gustafson RL, Paleos J (1971) Interactions responsible for the selective adsorption oforganics on organic surfaces. In: Faust SJ, Hunter JV (eds) Organic compounds inaquatic environments. Marcel Dekker, New York, pp 213–237

39. Riley JP, Taylor D (1969) Anal Chim Acta 46:30740. Osterroht C (1972) Kiel. Meeresforsch 28:4841. Osterroht C (1974) J Chromatogr 101:28942. Dravnieks A, Krotoszynski BK, Whitfield J, O’Donnell A, Burgwald T (1971) Environ

Sci Technol 5:122043. Kaiser R (1972) Haus Tech Essen Vortragsveroeff 1972:1344. Zlatkis A, Lichtenstein HA, Tichbee A (1973) Chromatographia 6:6745. Bergert KH, Betz V, Pruggmayer D (1974) Chromatographia 7:11546. Bertsch W, Chang RC, Zlatkis A (1974) J Chromatog Sci 12:17547. Bertsch W, Anderson E, Hylser G (1975) J Chromatogr 112:70148. Novak J, Zluticky J, Kubelka V, Mostecky J (1973) J Chromatogr 76:4549. Wasik SP (1974) J Chromatog Sci 12:84550. Kuo PRK, Chian ESK, DeWalle FB, Kim JH (1977) Anal Chem 49:102351. Colenut BA, Thorburn S (1980) Int J Environ Anal Chem 7:23152. Colenut BA, Thorburn S (1980) Int J Environ Anal Chem 7:2553. Grob K (1973) J Chromatogr 84:25554. Grob K, Grob G (1974) J Chromatogr 90:30355. Grob K, Grob G (1975) J Chromatogr 106:29956. Westerdorf RG (1982) Int Lab September:3257. Waggott A, Reid WI (1977) Separation of non-volatile organic carbon in sewage effluent

using high-performance liquid chromatography, pt 1. Technical report no. 52. WaterResearch Centre, Stevenage, UK

58. Corwen JF (1970) Volatile organic materials in sea water. In: Hood W (ed) Organicmatter in natural waters. Publication no. 1. Institute of Marine Science, University ofAlaska, pp 169–180

59. Kaiser KLE, Oliver BG (1976) Anal Chem 48:220760. Gjavotchanoff S, Luessen H, Schlimme E (1971) Gas und Wasser 112:44861. Chesler SN, Gump BH, Hertz HP, May WE, Wise SA (1978) Anal Chem 50:80562. Heyndrickx A, Peteghem C (1975) Eur J Toxicol 8:27563. Dowty B, Green L, Laseter JL (1976) J Chromatog Sci 14:18764. Waggott A (1978) Proceedings of the Analytical Division of the Chemical Society 15:23265. Bellar TA, Lichtenberg JJ (1974) J Amer Water Works Assoc 566:73966. Friant SE, Suffet IH (1979) Anal Chem 51:216167. Malter L, Vreden V (1978) Forum Stadte-Hygiene 29:3768. Cowen WF, Cooper WJ, Hisfill JW (1975) Anal Chem 47:248369. Chian ESK, Kuo PPK, Cooper WJ, Cowen WF, Fuentes RC (1977) Environ Sci Technol

11:28270. Luessem H (1972) Haus Tech Essen Vortragveroeff 1972:6971. Sheldon RW (1972) Limnol Oceanog 17:49472. Gjesing E, Lee FG (1967) Environ Sci Technol 1:63173. Sirotkina IS, Varshal GM, Lur’e YY, Stepanova NP (1974) Zhur Anal Khim 29:162674. Pietrzyk DI, Chu CH (1977) Anal Chem 49:75775. Pietrzyk DI, Chu CH (1977) Anal Chem 49:86076. Scott RPW, Kucera P (1973) Anal Chem 45:74977. Snyder LR (1974) Anal Chem 46:1384


78. Engelhardt H (1975) Hochdruck Flüssigkeits–Chromatographie. Springer, Berlin, Ger-many

79. Ahling B, Jensen S (1970) Anal Chem 42:148380. Uthe JF, Reinke J (1972) Environ Lett 3:11781. Aue WA, Kapila S, Hastings CR (1972) J Chromatogr 73:9982. Derenbach JR, Ehrhardt M, Osterroht C, Petrick G (1978) Mar Chem 6:35183. Hayase K, Shitashima K, Tsubota H (1985) J Chromatogr 322:35884. Beroza M (1962) Anal Chem 34:180185. Rahn PC, Siggia S (1973) Anal Chem 45:233686. Williams PM (1968) Nature 219:15287. Williams PM, Gordon LI (1970) Deep Sea Res 17:19

9 Organic Compounds

While there has always been some interest in the nature of the organic com-pounds in seawater, identification of actual compounds has progressed slowlybecause of the low concentrations present. With a total organic carbon concen-tration of 0.5–1.5 mg/l of carbon, the total concentration of any single organiccompound is likely to be less than 10–7 M. Therefore, in the past, identificationof individual compounds has been limited to those few for which specific,sensitive chemical methods existed. These methods were usually spectropho-tometric, and were often developments of methods originally used in clinicalchemistry.

The advent of the newer physical methods of separation and identification,together with the impetus given to the field by the imposition of anti-pollutionlegislation, has resulted in a flood of new and often unproven methods. Whilemost of these methods were specifically designed to measure materials addedto the environment by man’s activities, in many cases they have added greatlyto our knowledge of the naturally occurring compounds as well.

Until the advent of modern instrumental methods of analysis, the best wecould hope to do was to measure the amounts of certain broad classes ofcompounds present, as, for example, the total protein or total carbohydrate.

Using the newer methods, such as gas chromatography, liquid–liquid chro-matography, fluorometry, and mass spectrometry, it is possible to measuremany compounds at the parts-per-billion level, and a few selected compoundswith special characteristics at the parts-per-trillion level. Even with these sen-sitivities, however, a considerable concentration must usually be undertakento permit the chemical or physical fractionation necessary to render the finalanalyses interpretable. A major effort has therefore been expended on the studyof methods of separation and concentration, and this is discussed further inChap. 8.

A problem which has been less well recognised is that of contaminationin sampling and sample handling. The oceanographic vessel itself is a majorsource of contamination; from the moment that the ship stops on station,a surface film of oil, flakes of metal and rust spreads out in all directions.The means of sampling also often acts as a source of contamination. Once thesample is on board, along with the normal problems of contamination throughsample handling and through high blank values from the reagents used, we

366 9 Organic Compounds

must also attempt to cope with chemical and biological changes occurringduring storage, since many of the modern instruments are not easily takento sea. In environmental chemistry, the work of Analyst does not begin withthe delivery of the sample to the laboratory; every aspect of sampling, storage,and pretreatment must be considered part of the analyst’s domain. This isdiscussed further in Chap. 1.

Hydrocarbons

Probably the most studied group of organic compounds in seawater is thehydrocarbons, not because of the importance of the naturally occurring ma-terials, but because of the continuing threat of large-scale pollution. However,the methods devised for the measurement of anthropogenic hydrocarbons willalso measure the natural materials. A major problem, not altogether solved, isthat of distinguishing between the two sources.

There have been many reviews of analytical methods for hydrocarbons,particularly those given in [1–11].

Burgess [538] has reviewed the charactisation and identification of organictoxicants in marine waters

9.1Aliphatic Hydrocarbons


Fam et al. [11] determined hydrocarbons in run-off water from catchmentsin San Francisco Bay using liquid chromatography and high-resolution gaschromatography.

Wade and Quinn [12] measured the hydrocarbon content of sea surfaceand subsurface samples. Hydrocarbons were extracted from the samples andanalysed by thin-layer and gas–liquid chromatography. The hydrocarbon con-tent of the surface micro layer samples ranged from 14 to 599 µg/l with anaverage of 155 µg/l, and the concentration in the subsurface samples rangedfrom 13 to 239 µg/l and averaged 73 µg/l. Several isolated hydrocarbon frac-tions were analysed by infrared spectrometry and each fraction was found tocontain a minimum of 95% hydrocarbon material, including both alkenes andaromatics.

9.1.2Dynamic Headspace Analysis

May et al. [13] have described a gas chromatographic method for analysinghydrocarbons in marine sediments and seawater which is sensitive at the

9.1 Aliphatic Hydrocarbons 367

sub microgram per kilogram level. Dynamic headspace sampling for volatilehydrocarbon components, followed by coupled-column chromatography foranalysing the nonvolatile components, requires minimal sample handling, thusreducing the risk of sample component loss and/or sample contamination.The volatile components are concentrated on a Tenax gas chromatographicprecolumn and determined by gas chromatography or gas chromatography–mass spectrometry.

Otherworkershavediscussed theapplicationofdynamicheadspaceanalysisto the determination of aliphatic hydrocarbons in seawater [14–18].

A second system, the removal of volatiles by vacuum, can be set up in twoways; either as a flow-through or as a batch process. As a flow-through process,the sample is drawn continuously through the system, and the gases taken offby the vacuum pass through a sampling loop. Periodically, the material in theloop is injected into the gas chromatograph. In this manner it is possible toderive almost continuous profiles of volatile hydrocarbon concentrations [19].

In the batch mode, a larger sample can be treated over a longer period,and the volatiles collected by cold-trapping or adsorption. These techniquesare not as fast as flow-through sampling, nor do they permit semi-continuousprofiling, but they result in greater concentrations of the hydrocarbons, andthus in greater sensitivity [20].

The third technique, stripping-out, is by far the most common. In this tech-nique, an inert gas is bubbled through the sample to remove the volatile ma-terials. When the concentration of hydrocarbons is great enough, as, perhaps,after a petroleum spill, the emergent gas stream can be sampled directly [21].This is seldom the case in true oceanic samples, however, and some form ofconcentration is needed.

It is possible to collect the volatiles in a cold trap [22]. A more favouredtechnique is the collection of the gases by adsorption on some support suchas one of the Chromosorbs, or Tenax GC [13, 22, 23]. The volatiles are thendesorbed by heating and injected into a gas chromatograph.

Of the three general methods, the last seems to be the most practical. Theo-retically, with high enough concentrations of hydrocarbons, the first method,the headspace analysis, should be both the most accurate and the easiest to cal-ibrate. Operationally, it leaves much to be desired both because of the problemsof sensitivity and those of the accommodation of the larger molecules in water.The second method, vacuum degassing, requires much more equipment thanthe third method and requires that large amounts of water vapor be removedbefore the sample is injected into the gas chromatograph. The last method isso much less complicated that even with its calibration problems it has beenadopted almost universally.

While the gases used in stripping are usually air, nitrogen, or helium,electrolytically evolved hydrogen has been used as a collector for hydrocar-bons [24]. In this technique the gas is not passed through a column of adsor-bent, but instead collects in the headspace of the container. Since the volume


of seawater and of hydrogen is known, the hydrocarbon concentration in theheadspace can be used to calculate the partition coefficients and the concen-trations of hydrocarbon in the seawater. This method is capable of determining1 µg/l of volatile hydrocarbons in seawater.

9.1.3Raman Spectroscopy

Ahmadyian and Brown [25] have used laser Raman spectroscopy to identifypetroleums.

9.1.4Flow Calorimetry

Zsolnay and Kiel [26] have used flow calorimetry to determine total hydro-carbons in seawater. In this method the seawater (1 litre) was extracted withtrichlorotrifluoroethane (10 ml) and the extractwas concentrated,first inavac-uum desiccator, then with a stream of nitrogen to 10 µl. A 50 µl portion of thissolution was injected into a stainless steel column (5 cm × 1.8 mm) packedwith silica gel (0.063–0.2 mm) deactivated with 10% of water. Elution was ef-fected, under pressure of helium, with trichlorotrifluoroethane at 5.2 ml perhour and the eluate passed through the calorimeter. In this the solution flowedover a reference thermistor and thence over a detector thermistor. The latterwas embedded in porous glass beads on which the solutes were adsorbed withevolution of heat. The difference in temperature between the two thermistorswas recorded. The area of the desorption peak was proportional to the amountof solute present.

9.2Aromatic Hydrocarbons


Booksh et al. [27] employed an excitation/emission matrix imaging spectroflu-orometer for quantitation of two fluorescent compounds, naphthalene andstyrene, contained in ocean water exposed to gasoline. Multidimensional par-allel factor (PARAFAC) analysis models were used to resolve the naphthaleneand styrene fluorescence spectra from a complex background signal and over-lapping spectral interferents not included in the calibration set. Linearity wasdemonstrated over 2 orders of magnitude for determination of naphthalenewith a detection limit of 8 parts per billion. Similarly, nearly 2 orders of mag-nitude of linearity were demonstrated in the determination of styrene with an

9.3 Polyaromatic Hydrocarbons 369

11 ppb limit of detection. Furthermore, the synthesis of the EEM spectroflu-orometer and the PARAFAC analysis for unbiased prediction of naphthaleneand styrene concentration in mixture samples containing uncalibrated spectralinterferents was demonstrated.

9.2.2High-Performance Liquid Chromatography (HPLC)

By its very nature, the gas chromatograph is only useful with those compoundswhich can be made volatile in some manner. Compounds that are nonvolatile atthe temperatures that can be achieved in the gas chromatograph injection port,or those that degrade and polymerise, will be left as a residue in the injectionportorat the topof the column.For these compounds,high-performance liquidchromatography is the natural technique. The weak point of this techniquehas been the sensitivity of the detectors; the common commercially availabledetectorsmeasure refractive index, and lightabsorptionandfluorescence in theultraviolet and visible. Of these, only the fluorescence detector can approachthe sensitivity of the gas chromatograph detectors, and it is useful only forthose few compounds that are naturally fluorescent. There have also beenattempts to link the liquid chromatograph to flame ionisation detectors andatomic absorption spectrometers.

HPLC has been used, with an ultraviolet absorption detector set for 254 nm,for the determination of aromatic hydrocarbons and with a flow calorimeterfor the detection of all hydrocarbons. Increased sensitivity and decreased inter-ference can be achieved with the ultraviolet absorption detector by measuringabsorption at two wavelengths and using the ratios of the absorption at thosewavelengths [28].

9.3Polyaromatic Hydrocarbons

Hiltabrand [29] has investigated the fluorometric determination of polyaro-matic aromatic hydrocarbons in seawater.

Payne [30] carried out a field investigation of benzopyrene hydrolysateinduction as monitor for marine petroleum pollution. Isaaq et al. [31] isolatedstable mutagenic ultraviolet photodecomposition products of benzo(a)pyreneby thin-layer chromatography.

Fuoco et al. [539] has reported the analysis of priority pollutants in seawaterusing online supercritical fluid chromatography, cryotrap gas chromatogra-phy–mass spectrometry.Using this systempolynuclear aromatichydrocarbonsand polychlorobiphenyls were measured in seawater with recoveries betterthan 75%.

Law et al. [540] have recently reviewed methods for the analysis of polyaro-matic hydrocarbons in marine water.


9.4Oil Spills

In the early days of pollution research, many methods were investigated in thehope of finding a single technique which would infallibly link the deed andthe doer, the oil spill and the leaking container. The research was aided bya number of cases in which the provenance was obvious; the Torrey Canyonand the Arrow incidents are two examples. Thus it was possible to study thechanges brought about by weathering and to discover how these changes wouldhamper identificationof the sourceof the spill. Thedifficultyof attempting suchidentification using only a single technique was clearly shown by the numberof proposals to add radioactive or chemical labels to petroleum products asthey were loaded into the tankers.

Of the methods developed for the identification of hydrocarbon mixtures,only coupled gas chromatography–mass spectrometry holds any real promiseof certain identification and this only at a prohibitive cost in time spent charac-terising minor peaks. It would be far more efficient to develop rapid screeningprocedures which would eliminate all but a few possibilities, and then use gaschromatography–mass spectrometry to isolate and identify a few key peaks toconfirm the characterisation. This is precisely the scheme adopted indepen-dently by a number of laboratories.


This method was originally used to detect oil in surveys of oil in seawater(Zitko and Carson [32]; Michalik and Gordon [33]; Levy [34, 35]; Levy [36]).

Van Duuren [37] examined the use of emission and excitation spectra in theidentification of aromatic hyorocarbons. Contour diagrams of fluorescenceactivity at various excitation and emission wavelengths have been used asa means of identifying petroleum residues.

However, the main use of fluorescence has been in the semi-quantitativedetermination of aromatic hydrocarbons by extraction into an organic solvent,followed by excitation at a standard wavelength and comparison with theemission from a chosen standard. These techniques have been studied bymany workers [38–42].

The difficulties in the use of fluorescence for quantitative measurement ofhydrocarbons are much like those for the ultraviolet absorption methods. Eachcompound has its own excitation and emission maxima, with the fluorescencequantum yields varying sometimes by an order of magnitude. Thus the amountof hydrocarbon reported by an analysis will depend upon the emission and ex-citation wavelengths chosen, and upon the compound selected as the standard.

Petroleum products contain many fluorescing compounds, e.g., aromatichydrocarbons, polycyclic aromatic hydrocarbons, and various heterocyclic

9.4 Oil Spills 371

compounds. The use of fluorescence technique and instrumentation has ledto the use of this technique for the identification of crude and residual oilpollutants in a marine environment [43, 44] and of motor oils and relatedpetroleum products [45–48].

Maher [49] used fluorescence spectroscopy for monitoring petroleum hy-drocarbon contamination in estuarine and ocean waters.

An ingenious variation on the standard fluorescence methods was proposedby Red’kin et al. [50]. Water samples were extracted with non-polar solvents,transferred into hexane and the hexane solution frozen at 77 K. At that temper-ature the normally diffuse luminescence emission bands are present as sharpemission lines, making identification of fluorescing compounds considerablysimpler. In the case of a complex mixture, some separation by column or thinlayer chromatography might be necessary.

Second-order and fourth-order derivative synchronous spectrometry hasbeen used to fingerprint crude oil and fuel oil spills in seawater [51].

9.4.2Infrared Spectroscopy

Kawahara [52–54] has described an infrared spectroscopic method applicableto essentially nonvolatile petroleum products.

Heavy residual fuel oils and asphalts are not amenable to gas chromatogra-phy and give similar infrared spectra. However, a differentiation can be madeby comparing certain absorption intensities [52]. Samples were extracted withchloroform, filtered, dried, and the solvent evaporated off at 100 ◦C for a fewminutes using an infrared lamp. A rock salt smear was prepared from theresidue in a little chloroform, and the final traces of solvent removed usingthe infrared lamp. The method, which in effect compares the paraffinic andaromatic nature of the sample, involves calculation of the following absorptionintensity ratios:

• (13.88 µm polymethylene chain)/(7.27 µm methyl groups)• (3.28 µm aromatic C–H)/(3.42 µm aliphatic C–H)• (12.34 µm aromatic rings)/(7.27 µm methyl groups)• (6.25 µm aromatic C–C)/(7.27 µm methyl groups)• (12.34 µm aromatic rings)/(13.88 µm polymethylene chain)• (6.25 µm aromatic C–C)/(13.88 µm polymethylene chain)

Peaks observed at 5.90 µm and 8.70 µm were thought to reflect oxidative effectson the asphaltic material, while asphaltic sulfoxide and sulfone were tentativelyinferred from bands at 9.76, 8.66, and 7.72 µm. The 12.34/13.88, 12.34/7.27,and 6.25/13.88 µm ratios tended to show the greatest difference between dif-ferent samples. When the ratio 12.34/7.27 µm versus 12.34/13.88 µm were plot-ted graphically, the intermediate fuel oils behaved similarly [53]. Weatheringcaused fuel oils to fall below the curve although with asphalts the effect was


significantly less. Since no prior purification was employed the method relieson an uncontaminated, unweathered sample of oil being available.

Mattson and Mark [55,56] reported some criticism of Kawahara’s technique.They claim that evaporation of the solvent chloroform by infrared heating re-moves volatiles and causes large changes in the ratios. An oil sample was shownto suffer such alteration by the infrared during repeated analysis. The absorp-tion of all bands decreased nonuniformly between 20 and 100% over a periodof 30 min. They propose the application of internal reflection spectrometry asa rapid, direct qualitative technique requiring no sample pretreatment.

In contrast to infrared spectrometry there is no decrease in relative sen-sitivity in the lower energy region of the spectrum, and since no solvent isrequired, no part of the spectrum contains solvent absorptions. Oil samplescontaminated with sand, sediment, and other solid substances have been anal-ysed directly, after being placed between 0.5 mm 23-reflection crystals. Crudeoils, which were relatively uncontaminated and needed less sensitivity, weresmeared on a 2 mm 5-reflection crystal. The technique has been used to differ-entiate between crude oils from natural marine seepage, and accidental leaksfrom a drilling platform. The technique overcomes some of the faults of in-frared spectroscopy, but is still affected by weathering and contamination ofsamples by other organic matter. The absorption bands shown in Table 9.1 areimportant in petroleum product identification.

Kawahara and Ballinger [53, 57] has used their method to characterisea number of known and unknown petroleum samples. All of these studiesused the normal transmission method to obtain infrared spectra; however,the feasibility of using internal reflection to obtain infrared spectra has beendemonstrated by several groups (Mattson and Mark [55], Mark et al. [58],

Table 9.1. Absorption bands important in petroleum products identification

Band (µm) Compound

3.23–2.78 Water3.28 Aromatic CH3.29 – CH33.42 > CH23.51 > CH25.88 > C = O6.25 Aromatic C–C6.90 > CH27.27 – CH39.71 > S=O, PO411.63 Aromatic CH13.60 Aromatic CH13.98 Long chain –CH2–


9.4 Oil Spills 373

Baier [59]). The advantage of the latter method is that chemical extraction ofpetroleum from such as sand and water is unnecessary.

Pierre [60] has reported a study of the characterisation of the surface of oilslicks by infrared reflective spectroscopy. A double-beam spectrophotometerwas modified for studying the reflectance spectra (at angles of incidence 45◦,60◦, 70◦) of oil layers (20–30 µm thick) on the surface of water using purewater as reference.

Various other workers have discussed the application of this technique tooil spill analysis [61–63].


Various workers have discussed this technique [43, 64–72].Ramsdale and Wilkinson [66] have identified petroleum sources of beach

pollution by gas chromatography. Samples containing up to 90% of sandor up to 80% of emulsified water were identified, without pretreatment bygas chromatography, on one of a pair of matched stainless steel columns(750 × 3.2 mm id) fitted with precolumns (100 mm) to retain material of highmolecular weight, the second column being used as a blank. The columnpacking is 5% of silicone E 301 on Celite (52–60 mesh), the temperature isprogrammed at 5 ◦C per minute from 50 ◦C to 300 ◦C, nitrogen was used ascarrier gas, and twin flame ionisation detectors were used.

Adlard et al. [74] improved the method of Ramsdale and Wilkinson [66]by using an S-selective flame photometric detector in parallel with the flameionisation detector. Obtaining two independent chromatograms in this waygreatly assists identification of a sample. Evaporative weathering of the oilsamples has less effect on the information attainable by flame photometricdetection than on that attainable by flame ionisation detection. A stainlesssteel column (1 m × 3 mm id) packed with 3% of OV-1 on AWDMCS Chro-mosorb G (85–100 mesh) was used, temperature programmed from 60 ◦C to295 ◦C per minute with helium (35 ml/min) as carrier gas, but the utility of thetwo-detector system is enhanced if it is used in conjunction with a stainlesssteel capillary column (20 × 0.25 mm) coated with OV-101 and temperatureprogrammed from 60 ◦C to 300 ◦C at 5◦ per minute, because of the greaterdetail shown by the chromatograms.

Brunnock et al. [67] have also determined beach pollutants. They showedthat weathered crude oil, crude oil sludge, and fuel oil can be differentiatedby the n-paraffin profile as shown by gas chromatography, wax content, waxmelting point, and asphaltene content. The effects of weathering at sea on crudeoil were studied; parameters unaffected by evaporation and exposure are thecontents of vanadium, nickel, and n-paraffins. The scheme developed for theidentification of certain weathered crude oils includes the determination ofthese constituents, together with the sulfur content of the sample.


Adlard and Matthews [75] applied the flame photometric sulfur detectorto pollution identification. A sample of the oil pollutant was submitted to gaschromatography on a stainless steel column (1 m × 3 mm) packed with 3% ofOV-1 on AWDMCS Chromosorb G (85–100 mesh). Helium was used as carriergas (35 ml/min) and the column temperature was programmed from 60 ◦Cto 295 ◦C at 5◦ per minute. The column effluent was split between a flameionisation and a flame photometric detector. Adlard and Matthews [75] claimthat the origin of oil pollutants can be deduced from the two chromatograms.The method can also be used to measure the degree of weathering of oilsamples.

Boylan and Tripp [76] determined hydrocarbons in seawater extracts ofcrude oil and crude oil fractions. Samples of polluted seawater and the aqueousphases of simulated samples (prepared by agitation of oil–kerosene mixturesand unpolluted seawater to various degrees) were extracted with pentane.Each extract was subjected to gas chromatography on a column (8 ft × 0.06 in)packed with 0.2% of Apiezon L on glass beads (80–100 mesh) and temper-atures programmed from 60 ◦C to 220 ◦C at 4 ◦C per minute. The compo-nents were identified by means of ultraviolet and mass spectra. Polar aro-matic compounds in the samples were extracted with methanol-dichlorome-thane (1:3).

Investigations on pelagic tar in the North West Atlantic have been car-ried out using gas chromatography [77]. This report collects together theresults of various preliminary investigations. It is in the Sargasso Sea wherethe highest concentrations (2–40 mg/m2) occur, and on beaches of isolatedislands, such as Bermuda. These workers discuss the occurrence, structure,possible sources, and possible fate of tar lumps found on the surface of theocean.

Zafiron and Oliver [78] have developed a method for characterising envi-ronmental hydrocarbons using gas chromatography. Solutions of samples con-taining oil were separated on an open-tubular column (50 ft × 0.02 in) coatedwith OV-101 and temperature programmed from 75 ◦C to 275 ◦C at 6 ◦C perminute; helium (50 ml/min) was used as carrier gas and detection was by flameionisation. To prevent contamination of the columns from sample residues thesample was injected into a glass-lined injector assembly, operated at 175 ◦C,from which gases passed into a splitter before entering the column. Analysis ofan oil on three columns gave signal intensity ratios similar enough for directcomparison or for comparison with a standard. The method was adequate forcorrelating artificially weathered oils with sources and for differentiating mostof 30 oils found in a sea port.

Garra and Muth [80] and Wasik and Brown [81] characterised crude, semi-refined, and refined oils by gas chromatography. Separation followed by dual-response detection (flame ionisation for hydrocarbons and flame photometricdetection for S-containing compounds) was used as a basis for identifyingoil samples. By examination of chromatograms, it was shown that refinery

9.4 Oil Spills 375

oils can be artificially weathered so that the source of the oils can be deter-mined.

Hertz et al. [79] have discussed the methodology for the quantitative andqualitative assessment of oil spills. They describe an integrated chromato-graphic technique for studies of oil spills. Dynamic headspace sampling,gas chromatography, and coupled-column liquid chromatography are usedto quantify petroleum-containing samples, and the individual components inthese samples are identified by gas chromatography and mass spectrometry.

Rasmussen [82] describes a gas chromatographic analysis and a methodfor data interpretation that he has successfully used to identify crude oil andbunker fuel spills. Samples were analysed using a Dexsil-300 support coatedopen tube (SCOT) column and a flame ionisation detector. The high-resolutionchromatogram was mathematically treated to give “GC patterns” that werea characteristic of the oil and were relatively unaffected by moderate weath-ering. He compiled the “GC patterns” of 20 crude oils. Rasmussen [82] usesmetal and sulfur determinations and infrared spectroscopy to complement thecapillary gas chromatographic technique.

The gas chromatograms of most oil samples examined had similar basicfeatures. All were dominated by the n-paraffins, with as many as 13 resolvedbut unidentified smaller peaks appearing between the n-paraffin peaks ofadjacent carbon numbers. Each oil had the same basic peaks, but their relativesize within bands of one carbon number varied significantly with crude source.

9.4.4Gas Chromatography–Mass Spectrometry (GC–MS)

In some cases it is necessary unambiguously to identify selected componentsseparated during gas chromatographic examination of oil spill material. Suchmethods are needed from the standpoint of the enforcement of pollutioncontrol laws. The coupling of a mass spectrometer to the separated componentsemerging fromagas chromatographic separationcolumnenables suchpositiveidentifications to be made.

Smith [83] classified large sets of hydrocarbon oil infrared spectral databy computer into “correlation sets” for individual classes of compounds. Thecorrelation sets were then used to determine the class to which an unknowncompound belongs from its mass spectral parameters. A correlation set isconstructed by use of an ion-source summation, in which a low resolutionmass spectrum is expressed as a set of numbers representing the contributionto the total ionisation of each of 14 ion series. The technique is particularlyvaluable in the examination of results from coupled gas chromatography–massspectrometry of complex organic mixtures.

Walker et al. [84] examined several methods and solvents for use in theextraction of petroleum hydrocarbons from estuarine water and sediments,during an in situ study of petroleum degradation in sea water. The use of


hexane, benzene, and chloroform as solvents is discussed and compared, andquantitative and qualitative differences were determined by analysis usinglow-resolution computerised mass spectrometry. Using these data, and dataobtained following the total recovery of petroleum hydrocarbons, it is con-cluded that benzene or benzene-methanol azeotrope are the most effectivesolvents.

Brown and Huffman [85] reported an investigation of the concentrationand composition of nonvolatile hydrocarbons in Atlantic Ocean and nearbywaters. Sea water samples were taken at depths of 1 and 10 m and the non-volatile hydrocarbons were identified by mass spectrometric techniques. Theresults show that the nonvolatile hydrocarbons in Atlantic and nearby wa-ters contained aromatics at lower concentrations than would be expected ifthe source of the hydrocarbons were crude oil or petroleum refinery prod-ucts. Hydrocarbons appeared to persist in the water to varying degrees, withthe most persistent being the cycloparaffins, then isoparaffins, and finally thearomatics.

Albaiges and Albrecht [86] propose that a series of petroleum hydrocar-bons of geochemical significance (biological, markers) such as C20–C40 acyclicisoprenoids and C27 steranes and triterpenes be used as passive tags for thecharacterisation of oils in the marine environment. They use mass fragmentog-raphy of samples to make evident these series of components without resort-ing to complex enrichment treatments. They point out that computerised gaschromatography–mass spectrometry permits multiple fingerprinting from thesame gas chromographic run. Hence rapid and effective comparisons betweensamples and long-term storage of the results for future examination can becarried out.

Another relevant feature of the gas chromographic profile is the acyclicisoprenoid hydrocarbon pattern that is made evident with capillary columnsor by the inclusion of the saturated fraction in 5 Å (0.5 nm) molecular sieves orin urea. The predominant peaks usually correspond to the C19 (pristaine) andC20 (phytane) isomers, which ratios serve as an identification parameter [87],although the series extends to lower and higher homologues.

The sulfur compounds that are present in minor quantities in petroleumproducts also exhibit a typical gas chromatographic fingerprint easily ob-tained by flame photometric detection. This fingerprint has been introducedto complement the flame ionisation detection chromatogram with the aim ofresolving the ambiguities or increasing the reliability in the identification ofthe pollutants [74].

All the above fingerprints exhibit a different usefulness for characterisingoils. Their variation between crudes and their resistance to the sea weather-ing processes are not enough, in many cases, for providing the unequivocalidentification of the pollutant. The n-paraffins can, apparently, be removedby biodegradation, as can the lower acyclic isoprenoids at respectively slowerrates.

9.5 Carboxylic Acids and Hydroxy Acids 377

9.4.5Miscellaneous

Other techniques that have been used in the examination of hydrocarbonsin seawater resulting from oil spills include gel permeation chromatogra-phy [88], mass spectrometetry [85], turbidimetry [89, 90], and paper chro-matography [91].

Characterisations of crude oil based on their metal [65,92], nitrogen [92,93],and sulfur [99] contents have been carried out.

9.5Carboxylic Acids and Hydroxy Acids


If ameasurementof total fatty acid concentration isdesired, short of attemptingto sum the amounts of each compound found by gas chromatography, someindirect method must be employed. Harwood and Huyser [94] made iron (III)hydroxamates of the fatty acids and measured these colorimetrically.

Antia et al. [96] proposed a colorimetric method comprising a concen-tration step and a reaction with 2,7-dihydroxynaphthalene. This method hasa detection limit of 0.1 mg/l. Methods comprising concentration by adsorptionon alumina, followed by elution and colour development, yielded a detectionlimit of 5 ng/l [97, 98]. These methods have not been too successful in handsother than those of the original authors, possibly because of the amount ofmanipulation necessary in the analysis.

Stradomskaya and Goncharova [99] have developed an extraction andcolour development method for formic acid which should work in seawa-ter. The sample is acidified to pH 2 and extracted with diethyl ether. Afterremoval of the ether, the formic acid is reduced to formaldehyde and deter-mined spectrophotometrically with chromotropic acid. A sensitivity of 1 µg/lwith a normal range of 0.9 µg/l is claimed.


Trifluoroacetatehasbeendetermined inseawater inamountsdownto32 ng/lbyextraction with methyl tert butyl ether and derivatised with pentafluorophenyldiazomethane prior to gas chromatography mass spectrometry (negative ion-isation) [541].

One of the earliest applications of gas chromatography to marine problemswas in the measurement of fatty acids in seawater. In general, the gas chromato-graphic method has employed extraction into organic solvent, followed by the


formation of a volatile derivative, often the methyl ester. Garrett et al. [100]applied these techniques to materials collected from the surface layer, find-ing fatty acids from C8–C20 to be present. Slowey et al. [101] applied similaranalytical techniques to material isolated from the water column.

Earlier analytical results from the gas chromatographic analysis of fattyacids seemed to be very high. Williams [102] repeated much of the early work,using extreme care in the avoidance of contamination, and found very muchsmaller quantities. Papers concerned with the fatty acid content of the watercolumn and sediment include [103–109].

Analyses of the surface film have been performed [110–114] and analyticalmethods for the volatile members of the group have been reported [115–118].

Gas chromatography and gas chromatography–mass spectrometry havealso been used to measure fatty alcohol [119], phytol [120], and the severalisomers of inositol [121].

9.5.3Liquid Chromatography

All of the various chromatographic methods have been used for the separationand subsequent determination of the individual fatty acids. The earlier workused column and paper chromatography for the concentration and separa-tion of the acids, with spectrophotometric methods for the final measurement.Mueller et al. [122] concentrated carboxylic acids by evaporation and extrac-tion, then separated them by column and paper chromatography. Koyama andThompson [123] used vacuum distillation and extraction for the concentrationstep, and column chromatography for the separation. Maksimova and Pimen-ova [124], working with culture medium in which phytoplankton had beengrown, separated derivatives of the fatty acids on paper chromatograms. Thesemethods, although effective, are tedious; if they were to be attempted today,they would undoubtedly be adapted to thin-layer and liquid chromatography.

The fatty acidsmeasuredby these techniqueshave all been small monomericmolecules. Lamar and Goerlitz [125] studied the acidic materials in highlycoloured water and found that most of the nonvolatile material was composedof polymeric hydroxy carboxylic acids, with some aromatic and olefinic un-saturation. Their methods included gas, paper, and column chromatographywith infrared spectrophotometry as the major technique used for the actualcharacterisation of the compounds.

Horikawa [126] has adapted a thermal detector for the determination offormic, acetic, and propionic acids by liquid chromatography.

Gorcharova and Khomenko [127] have described a column chromato-graphic method for the determination of acetic, propionic, and butyric acidsin seawater, and thin-layer chromatographic methods for determining lactic,aconitic, malonic, oxalic, tartaric, citric, and malic acids. The pH of the sampleis adjusted to 8–9 with sodium hydroxide solution. It is then evaporated almost

9.5 Carboxylic Acids and Hydroxy Acids 379

to dryness at 50–60 ◦C and the residue washed on a filter paper with water acid-ified with hydrochloric acid. The pH of the resulting solution is adjusted to 2–3with hydrochloric acid (1:1), the organic acids are extracted into butanol, thenback-extracted into sodium hydroxide solution; this solution is concentratedto 0.5–0.7 ml, acidified and the organic acids separated on a chromatographiccolumn.

Application of high-performance liquid chromatography to the resolutionof complex mixtures of fatty acids [128, 129] has provided an alternative tothe high temperature separation obtained by gas chromatography. Both tech-niques have similar limits of detection, but lack the ability to analyse directlyenvironmental samples. Analysis requires that the fatty acids be separated fromthe organic and inorganic carbon matrices followed by concentration. Typi-cally, these processes can be accomplished simultaneously by the appropriatechoice of methods. Initial isolation of the fatty acids is based on the relativesolubility of the material of interest in an organic phase compared with theaqueous phase. Secondary separation is determined by the functional groupcontent and affinity for a solid support.

A coupled enzymic high-performance liquid chromatographic method hasbeen used to determine down to 0.1–0.8 µM of formate in sea water [130].

9.5.4Atomic Absorption Spectrometry (AAS)

Treguer et al. [95] determined total dissolved free fatty acids in sea water.The sample (1 litre) was shaken with chloroform (2 × 20 ml) to remove thefree fatty acids and the extract evaporated to dryness under reduced pres-sure at 50 ◦C. Chloroform-heptane (29:21) (2 ml) and fresh copper reagent(Mtriethanolamine-M acetic acetate-6.8% CuSO4–5H2O) solution (9:1:10)(0.5 ml) was added to this residue. The solution was shaken vigorously for3 min and centrifuged at 3000 rpm for 5 min. A portion (1.6 ml) of the organicphase was evaporated to dryness and 1% ammonium diethyldithiocarbamatesolution in isobutylmethyl ketone (2 ml) was added to the residue to forma yellow copper complex. The copper in the solution was determined by atomicabsorption spectrophotometry at 324.8 nm (air-acetylene flame). Palmitic acidwas used to prepare a calibration graph. The standard deviation for samplescontaining 30 µg/l of free fatty acids (as palmitic acid) was ±1 µg/l.

9.5.5Diffusion Method

Xiaoet al. [131]determinednanomolarquantitiesof individual low-molecular-weight carboxylic acids (and amines) in sea water. This method is based onthe diffusion of acids across a hydrophobic membrane to concentrate themand separate them from inorganic salts and most other dissolved organic


compounds. Acetic, propionic, butyric, valeric, pyruvic, acrylic, and benzoicacids were all found in measurable amounts in sea water.

9.6Ketones and Aldehydes

9.6.1Spectrophotometric Method, Fluorometric and Chemiluminescence Methods

A spectrophotometric method for aldehydes in either fresh or seawater was de-scribed by Kamata [132]. It used the colour-forming reaction between the alde-hyde, 3-methyl-2-benzothiazolone hydrazone, and ferric chloride, and claimeda sensitivity of 0.01 mg/l as formaldehyde equivalents. While Kamata clearlyfound evidence of the presence of aldehydes, the method appears to be notquite sensitive enough for the quantities to be found in seawater.

Eberhardt and Sieburth [133] also devised a spectrophotometric proce-dure for the determination of aldehydes in seawater. The method is based onthe reaction of aldehydes with 3-methyl-2-benzothiazolinone, hydrazone hy-drochloride, and ferric chloride to produce a coloured compound. A detectionlimit of 0.072 µMformaldehydeper litrewasobtainedusinga5 cmpath-length.

Automated colorimetric and fluorometric methods for aldehydes were pro-posed by Afghan et al. [134]. The colorimetric method used chromatotropicacid, while the fluorometric method was based on the reaction betweenformaldehyde, 2,4-pentadione, and ammonia. The sensitivities of these meth-ods were about the same as that of the Kamata method; they could be employedusefully in fresh water, but were marginal for seawater. A more sensitive ver-sion of the fluorometric method for formaldehyde was developed by Zika [135]using the reaction as described by Belman [136]. This version had a sensitivityof 1 ×10–7 M in sea water.

A continuous flow system utilising the oxidation of formaldehyde and gallicacid with alkaline hydrogen peroxide to produce a chemiluminescence wasstudied by Slawinska and Slawinski [137]. While the major peak of the chemi-luminescence spectrum occurred at 635 nm, the photomultiplier used summedall of the available light between 560 and 850 nm. The intensity of the chemi-luminescence was linearly proportional to formaldehyde concentration from10–7 to 10–2 M, producing a detection limit of 1 µg/l. This method should besensitive enough for use in seawater.

9.6.2Potential Sweep Voltammetry

Another fresh-water method which holds some promise for seawater analysisis twin cell potential sweep voltammetry, as proposed Afghan et al. [138].In this method, semicarbazones are formed by reaction with semicarbazide

9.7 Phenols 381

hydrochloride, and the peak height of the semicarbazones polarogram is thenmeasured. The shape of the polarographic peak is an indication of the type ofcompound present. The detection limit of this method is 1 ×10–8 M, and theeffective range runs from 10–8 to 10–4 M.


For those aldehydes and ketones which are volatile enough, gas chromatog-raphy of the headspace gases can be used, and this has been used to mea-sure acetone, butyraldehyde, and 2-butanone in oceanic waters. In a gaschromatography–mass spectrometry analysis of a single sample of volatilematerials concentrated from inshore waters onto Tenax GC, MacKinnon [139]reported tentative identification of methyl isopropyl ketone, bromoform, 4-methyl-2-pentanone, 2-hexanone, and 2-hexanal.

Corwen [140] used dynamic purge and trap analysis to determine ketonesand aldehydes (acetone, butyaldehyde, and 2-butanone) in seawater.

9.7Phenols


While the major emphasis in the analysis of phenols in seawater has beenon those compounds introduced by industrial processes, as much phenolicmaterial is probably added by the disintegration of fixed algae in the intertidalregions. A high value for total phenols, particularly in coastal waters, cannotbe interpreted simply as a high degree of industrial pollution; the kinds ofphenols present must also be ascertained.

A number of colorimetric methods for phenols in seawater have been re-ported. Alekseeva [141] oxidised the phenols to antipyrine dyes, then extractedthe dyes into an organic solvent of considerably smaller volume. The reporteddetection limit of the method was 20 µg/l, with a linear range of 20–80 µg/l.Goulden [142] used an automated system with steam distillation of the phe-nols from acidic solution, followed by formation of a coloured derivative andextraction of the derivative into an organic solvent for the final determination.Both 4-amino-antipyrine and 3-methyl-2-benzothiazolinone were evaluatedas reagents, with a detection limit of 0.2 µg/l for either reagent. Nitroani-line was used as the colorimetric reagent by Schlungbaum and Behling [143],while Stilinovic et al. [144] used both nitroaniline and 4-aminoantipyrine,and Gales [145] modified the 3-methylbenzothiazolinone method to lower thedetection limit to 1 µg/l.

Phenolic substances can be measured directly, without colour develop-ment, by the difference in their ultraviolet absorption in acidic and basic


solution [146]. Interference due to the ultraviolet absorption of non-ionisingnonphenolic organic species is cancelled out by this difference method. Themethod was adapted for natural fresh waters by Fountaine et al. [147] withthe use of two sealed hollow cathode lamps to monitor the difference spec-trum. Comparison with the 4-aminoantipyrine colorimetric method generallyshowed higher values for the ultraviolet difference method, probably due tothe presence of some para-blocked phenols, which will not react in the col-orimetric procedure. A simple photometric difference instrument has beendeveloped with a range of 100 µg/l full scale.


Boyd [148] determined ppq levels of phenols, cresols, and catechols in SanDiego Bay (CA, USA) water by aqueous acetylation of the sample followed bygas chromatography–mass spectrometry.

9.8Phthalate Esters

Among the most ubiquitous of man’s contributions to the environment arethe phthalate esters. These compounds are used extensively in the plasticsindustry and have been found in both fresh and seawater at concentrationsin the µg/l range. The usual method of analysis at this level of concentrationis gas chromatography; some form of concentration, usually adsorption ona column, is always needed, and great care is required to keep the backgroundlevels sufficiently low. Methods of collection, storage, and analysis have beendescribed [149–151]. The use of coupled gas chromatography–mass spectrom-etry as well as liquid chromatography was discussed by Hites [152].

9.9Carbohydrates

Of all the classes of organic material to be found in seawater, the carbohydratesare probably the most widely investigated. This is partly because of their rolein photosynthesis and their function as storage and structural compounds inalgae. Concentrations as low as a few µg/l are of significance.


The early methods were all spectrophotometric and were usually based onthe condensation of carbohydrates with a concentrated acid, usually sulfuric,followed by reaction of the condensed product with some colour-forming

9.9 Carbohydrates 383

compound including N-ethylcarbazole, anthrone [153], phenol, orcinol, andtryptophan [154].

Some comparative studies of these methods have been made: anthroneand N-ethylcarbazole were compared by Lewis and Rakestraw [153] and byCollier [155], and anthrone, phenol, orcinol and N-ethylcarbazole and L-tryp-tophan were examined by Josefsson et al. [156]. In general, the comparativestudies show anthrone to be more reliable than N-ethylcarbazole, althoughsomewhat less sensitive. However, Josefsson and co-workers [154] found thatof the five methods, the tryptophan method gave the best results when adaptedto automatic analysis, and was capable of analysis at concentrations of interestin seawater analysis, i.e., µg/l levels.

The acid condensation methods do not distinguish between monosaccha-rides and polysaccharides, as the various classes of carbohydrates each havedifferent absorption maxima, which results in different molar absorptions atany chosen wavelength. Furthermore, when treated with concentrated sulfuricacid, some three- and four-carbon compounds will condense into structureswhich will produce colours with those reagents. When the object of the analysisis to obtain some estimate of the total amount of carbohydrate or carbohydrate-like material present, the inclusiveness of these methods is useful. However,when the object is to distinguish between the easily metabolised simple sug-ars and the complex storage and structural materials, these methods give noinformation at all.

A spectrophotometric method that does not use condensation with sulfu-ric acid was proposed by Mopper and Gindler [157]. The method used thecopper (I) complex with 2,2-bicinchoninate to form a colour with simple sug-ars, with a hydrolysis step which is included to make simple sugars from thepolysaccharides. Thus the analysis of two aliquots, one of them hydrolysed,would yield values for both simple and combined sugars.

Another spectrophotometric method measuring both simple and combinedsugars was described in papers by Johnson and Sieburth [158] and Burney andSieburth [159]. The basic method comprised reduction of sugars to alditolswith sodium borohydride, and oxidation of the alditols to form free formalde-hyde. The formaldehyde was then determined spectrophotometrically with3-methyl-2-benzothiazolinone hydrazone hydrochloride.

In the determination of carbohydrates, sensitivity can often be increased byusing fluorescence rather than absorbance for the final determination. Withcompounds that arenotnormallyfluorescent, it becomesnecessary tofindfluo-rescent derivatives. Hirayama [160] concentrated the carbohydrates in coastalwater samples, using electrodialysis and evaporation, and made fluorescentderivatives using anthrone and 5-hydroxyl-1-tetralone, determining pentosesseparately from hexoses in the process. While this method does seem to havethe extra sensitivity expected from fluorescent methods, the extra manipula-tions render it unsatisfactory for routine use.


The methods discussed above are all class reactions, designed to estimatethe total amount of carbohydrate present and usually actually furnishing somesort of weighted average, weighted by the unequal responses of the differentclasses of sugars. There are a few methods that are specific for a single classof carbohydrates or for a single sugar. It has long been suspected that uronicacids make up a considerable portion of the dissolved organic carbon in theocean, but most of the carbohydrate methods do not measure these acids.Williams et al. [161] and Mopper [162] developed a modification of the acidicdecarboxylation method of Lefevre and Tollens, and found uronic acids inphytoplankton and in particulate organic carbon.

9.9.2Enzymic Methods

Enzymic methods are usually very specific and sensitive. Unfortunately theonly methods in the literature for carbohydrates are all for glucose. Hicks andCarey [163] reported such a method, with a fluorometric final measurement,which was down to 3 ×10–8 M. Andrews and Williams [164] used a preconcen-tration step, sorption onto charcoal, elution, and a final determination withglucose oxidase.

9.9.3Liquid Chromatography

Chromatographic techniques are required to distinguish between the differentclasses of carbohydrates. Chromatographic methods for both fresh and sea-water are described by Whittaker and Vallentyne [165], Degens et al. [166],and Starikova and Yablokova [167, 168]. These methods often include a hy-drolysis step to permit the measurement of the monomers held in polysac-charides. A discussion of these methods, together with a comparison be-tween paper chromatographic, colorimetric, and enzymatic methods can befound in Geller [169]. Josefsson [170] recommended a separation by liquidchromatography after desalting by ion exchange membrane electrodialysis.The electrodialysis cell used had a sample volume of 430 ml and an effectivemembrane-surface area of 52 cm2. Perinaplex A-20 and C-20 ion exchangemembranes were used. The water-cooled carbon electrodes were operated atup to 250 mA and 500 V. The desalting procedure normally took less than30 hours. After the desalting, the samples were evaporated nearly to dryness at40 ◦C in vacuo, then taken up in 2 ml of 85% ethanol, and the solution was sub-jected to chromatography on anion exchange resins (sulfate form) with 85%ethanol as mobile phase. By this procedure, it was possible to determine eightmonosaccharides in the range 0.15–46.5 µg/l with errors of less than 10%, andto detect traces of sorbose, fucose, sucrose, diethylene glycol, and glycerol inseawater.

9.9 Carbohydrates 385

Larsson and Degens [171] devised an automated system for determiningcarbohydrates in biological samples using partition chromatography for theseparation and the orcinol colorimetric method for the final analysis. Laterversions of this kind of autoanalyser, using tetrazolium blue or a CuI complexof bicinchoninate for the final measurement, have been reported [172].


The discovery of the usefulness of the trimethylsilyl derivatives for the gaschromatography of a sample was a major step forward in the analysis ofcomplex mixtures.

Such methods, usually also including a hydrolysis step to break downpolysaccharides, have been described by Modzeleski et al. [173] and Tesa-rik [174].

Eklund [175] developed a method for sensitive gas chromatographic anal-ysis of monosaccharides in seawater, using trifluoracetyl derivatisation andelectron capture detection. It is difficult to determine accurately the monosac-charide concentration by this method because a number of chromatographicpeaks result from each monosaccharide.

9.9.5Miscellaneous

Williams [176] has studied the rate of oxidation of C-labelled glucose in sea-water by persulfate. After the oxidation, carbon dioxide was blown off andresidual activity was measured. For glucose concentrations of 2000, 200, and20 µg/l, residual radioactivities (as percentage of total original radioactivity)were 0.04, 0.05, and 0.025, respectively, showing that biochemical compoundsare extensively oxidised by persulfate. With the exception of change of temper-ature, modifications of conditions had little or no effect. Oxidation for 2.5 h at100 ◦C was the most efficient.

While there appears to be a profusion of methods for the analysis ofcarbohydrates in seawater, a study of the actual capabilities of the meth-ods soon reveals that few of them furnish us with much useful informa-tion. At the moment only the methods of Johnson and Sieburth [158] andBurney and Sieburth [159], and the bicinchoninate method of Mopper andGindler [157], furnish any real estimate of the total amount of carbohy-drate present in seawater. For the analysis of the separate sugars, liquidchromatography of carbohydrate derivatives would seem to be the obvi-ous choice. Several methods of determining carbohydrates have been de-scribed [177–184].


9.10Cationic Surfactants


Wang and Pek [185] have described a simple titration method applicable to theanalysis of cationic surfactants in sea water. Methyl orange and azure A wereused as primary dye and secondary dye, respectively. The method is free frominterference by high levels of inorganic salts in sea water.


LeBihanandCourtot-Coupez [186] analysed freshwater for cationicdetergentsby a method based on atomic absorption spectrometry of the copper–detergentcomplex.


Hon-Nami and Hanya [187, 188] used chloroform extraction to extract linearalkyl benzene sulfates from seawater prior to analysis by GC–MS.

Riu et al. [542] has reported the determination of linear ethyl benzane-sulfates in coastal waters using automated solid-phase extraction followed bycapillary electrophoresis with ultraviolet detection, and confirmed by capil-lary electrophoresis-mass spectrometry. The detection limits were 1 µg/l when250 ml of coastal water was preconcentrated.

9.11Anionic Surfactants

There are two different classes of surface-active materials in seawater, thosethat are naturally present and those that have been added to the oceans byman’s activities. Most of the analytical methods proposed for use in seawateractually measure the anthropogenic input, and attempt as much as possibleto eliminate interferences from naturally occurring compounds. Yet sea foamwas known to exist long before detergents. It is to be expected that both kindsof surfactants would be concentrated at the air–sea interface.

9.11.1Titration

Titration methods are often used as alternatives to the spectrophotometricmethods. Wang et al. [189] determined anionic surfactants by adding an ex-

9.11 Anionic Surfactants 387

cess of cationic surfactant, acidifying, and then titrating with standard sodiumtetraphenylboron. The method was adapted for seawater by the same au-thors [190], and a simplified field kit was described.


The spectophotometric methylene blue method for anionic surfactants hasbeen applied to seawater. In one version, the surfactants are collected in ethylacetate. The solvent is then evaporated, the surfactants put back in solutionin water, and the standard spectrophotometric methylene blue method is ap-plied to this solution. In this manner, the salt error introduced by seawateris eliminated [195]. A similar method, with the methylene blue-surfactantcomplex extracted into chloroform, and measured directly was proposed byHagihara [192].

AmethodusingazureA insteadofmethylenebluehasbeenproposedbyDenTankelaar and Bergshoeff [193]. Workers at the Water Research Centre [194]described a methylene-blue based autoanalysis method for determining 0–1 µg/l anionic detergents in water and sewage effluents. This method wasbased on work of Longwell and Maniece [195], and subsequently modifiedby Abbott [196] and by Sodergren [197]. The Water Research Centre reportdescribes the method in detail and discusses its precision and accuracy.

Favretto and Tunis [198] extracted polyoxyethylene alkyl phenyl ethers aspicrates into an organic solvent, complexed the polymer with sodium ion, andmeasured the absorption complex. This is one of the few specific methodsavailable.

Bhat et al. [199] used complexation with the bis(ethylenediamine) cop-per (II) cation as the basis of a method for estimating anionic surfactantsin fresh estuarine and seawater samples. The complex is extracted into chlo-roform, and copper is measured spectrophotometrically in the extract using1,2(pyridyl azo)-2-naphthol. Using the same extraction system these workerswere able to improve the detection limit of the method to 5 µg/l (as linear alkylsulfonic acid) in fresh estuarine and seawater samples.


A rough estimate of the total amount of anionic surfactant present can beobtained by reacting the surfactant with a metal-containing material suchas bis(ethylenene diamine) copper (II) [199, 200, 203], or o-phenanthroline-CuSO4, extracting the complex into an organic solvent (209 MIBK), and deter-mining the metal by atomic absorption.

Le Bihan and Courtot-Coupez [202] used the copper complex and flamelessatomic absorption spectroscopy to determine anionic detergents. Crisp [200]


was the first to use the bis(ethylenediamine) CuII ion for the determination ofanionic detergents. These workers determine the concentration of detergentsby flame atomic absorption spectroscopy or by a colorimetric method. Thecolorimetric method was more sensitive, with a limit of detection of 0.03 µg/l(as linear alkyl sulfonic acid) compared with 0.06 µg/l for atomic absorptionspectroscopy. Crisp et al. [204] determined anionic detergents in fresh es-tuarine and seawater at the ppb level. The detergent anions in a 750 ml watersample are extracted with chloroform as an ion association compound with thebis(ethylene-diamine) copper (II) cation, and determined by atomic absorp-tion spectrometry using a graphite furnace atomiser. The limit of detection (aslinear alkyl sulfonic acids) is 2 µg/l.

Gagnon [203] has described a rapid and sensitive AAS method developedfrom the work of Crisp et al. [200] for the determination of anionic detergentsat the ppb level in natural waters. The method is based on determination byatomic absorption spectrometry using the bis(ethylene-diamine) copper (II)ion. The method is suitable for detergent concentrations up to 50 µg/l but itcan be extended up to 15 mg/l. The limit of detection is 0.3 µg/l.

The recovery of different concentrations of detergents added to seawater wasused to evaluate the accuracy of the method. The recovery is 80% at 1 µg/l butreaches 90% at 10 µg/l. The recovery is 97% or better at higher concentrations.Precision was very good.


The methylene blue reaction can also be used in a fractionation procedurefor surfactants. The complexes with methylene blue can be collected in anorganic solvent, concentrated, dissolved in methanol, and separated by high-performance liquid chromatography [205]. A variation of this method, per-mitting the collection of surfactant from large volumes of sample, should beworkable in seawater.

9.12Non-Ionic Surfactants


Favretto and co-workers [198, 206–208] have described direct spectrophoto-metric methods for non-ionic surfactants based on the formation of a sodiumpicrate surfactant adduct. This method has been applied to seawater. A meanvalue of 93±1% was obtained in recovery experiments on C12E9 (at an aqueousconcentration of 0.10 mg/l) extracted from synthetic sea water by means of this

9.12 Non-Ionic Surfactants 389

procedure. Therefore, a multiplication factor of 1.07 was adopted in correctingfor the extraction losses.


Courtot-Coupez and Le Bihan [209, 210] determined non-ionic detergents insea- and fresh-water samples at concentrations down to 2 µg/l ppm by ben-zene extraction of the tetrathiocyanatocobaltate (II) (NH4)2 (Co(SCN)4) [182]detergent ion-pair, followed by atomic absorption spectrophotometric deter-mination of cobalt [209].

Courtot-Coupez and Le Bihan [209,210] determined the optimum pH (7.4)for extraction of non-ionic surfactants with the above complex-benzene sys-tem. Cobalt in the extract is estimated by AAS after evaporation to dryness anddissolution of the residue in methyl isobutyl ketone. The method is applicableto surfactant concentrations in the range 0.02–0.5 mg/l and is not seriouslyaffected by the presence of anionic surfactants.

Crisp et al. [212] has described a method for the determination of non-ionicdetergent concentrations between 0.05 and 2 mg/l in fresh, estuarine, and sea-water based on solvent extraction of the detergent–potassium tetrathiocyana-tozincate (II) complex followed by determination of extracted zinc by atomicAAS. A method is described for the determination of non-ionic surfactants inthe concentration range 0.05–2 mg/l. Surfactant molecules are extracted into1,2-dichlorobenzene as a neutral adduct with potassium tetrathiocyanatozin-cate (II), and the determination is completed by AAS. With a 150 ml watersample the limit of detection is 0.03 mg/l (as Triton X-100). The method isrelatively free from interference by anionic surfactants; the presence of up to5 mg/l of anionic surfactant introduces an error of no more than 0.07 mg/l (asTriton X-100) in the apparent non-ionic surfactant concentration. The per-formance of this method in the presence of anionic surfactants is of specialimportance, since most natural samples which contain non-ionic surfactantsalso contain anionic surfactants. Soaps, such as sodium stearate, do not in-terfere with the recovery of Triton X-100 (1 mg/l) when present at the sameconcentration (i.e., mg/l). Cationic surfactants, however, form extractable non-association compounds with the tetrathiocyanatozincate ion and interfere withthe method.

9.12.3Liquid Chromatography–Mass Spectrometry (LC–MS)

Reversed phase liquid chromatography–mass spectrometry was applied toextracts of Jamaica Bay (New York) water to determine 1–300 µg/l amounts ofnonyl phenol ethoxylates and their metabolites [213].


9.13Aliphatic Chloro Compounds


Eklund et al. [214, 215] has developed a capillary column method [216, 217]for the determination of down to 1 µg/l volatile organohalides in waters thatcombines the resolving power of the glass capillary column with the sensi-tivity of the electron capture detector. The eluate from the column is mixedwith purge gas of the detector to minimise band broadening due to dead vol-umes. This and low column bleeding give enhanced sensitivity. Ten differentorganohalides were quantified in seawater. Using this technique these workersdetected bromoform in seawater for the first time.

Halogenated hydrocarbons in different waters were identified by compari-son with a standard solution. A chromatogram of a standard pentane solutionof various volatile organochlorine compounds including trihaloforms is shownin Fig. 9.1. Retention times were measured on two columns with different sta-tionary phases (SE-52 and Carbowax 400).

The approximate detection limits of the method range from 1 fg CCl4 to500 fg CH2CCl2.

9.13.2Purge and Trap Analysis

Kaiser and Oliver [218] applied dynamic purge and trap analysis to determinehalocarbons in seawater.

The Bellar et al. [219] purge and trap method has been applied to the deter-mination of vinyl chloride in seawater. Using the Hall electrolytic conductivitydetector, no response was obtained for the acetone used to prepare the vinylchloride standard solution.

Organic compounds are extracted from the charcoal with a small volume ofa suitable solvent such as carbon disulfide or dichloromethane, then collectedand injected into a capillary gas chromatograph or a capillary gas chromato-graph coupled with a mass spectrometer.

Grob et al. [220–223] have carried out very detailed and systematic studiesof the closed-loop gas stripping procedure and applied it to the determinationof µg/l of 1-chloroalkanes in water. Westerdorf [224] applied the technique tochlorinated organics, and aromatic and aliphatic hydrocarbons. Waggot andReid [225] reported that a factor of major concern in adapting the techniqueto more polluted samples is the capacity of the carbon filter, which usuallycontains only 1.5–2 mg carbon. They showed that the absolute capacity of sucha filter for a homologous series of 1-chloro-n-alkanes was 6 µg for completerecovery.

9.13 Aliphatic Chloro Compounds 391

Figure 9.1. Chromatogram of a standard mixture containing 11 organohalides. Stationaryphase, SE-52; 33 m × 0.3 mm id, injection temperature 200 ◦C, column temperature 50 ◦C,interface250 ◦C,detector temperature250 ◦C.Heliumcarriergasflowrate36 ml/s, scavengergas flow 30 ml/min. Split ratio 1.20. 1 CH2Cl2 2 CHCl3 3 CH3CCl2 4 CCl4 5 CHClCCl26 CHBrCl2 7 CBrCl3 8 CHBr2Cl 9 CCl2CCl2 10 CHCl2I 11 CHBr3. Source: [215]

9.13.3Head Space Analysis

Trichlorofluoromethane and dichlorofluoromethane have been determinedin seawater using headspace analysis and gas chromatography with electroncapture detection [226].


9.13.4Miscellaneous

Dawson et al. [227] described samplers for large-volume collection of seawater samples for chlorinated hydrocarbon analyses. The samplers use themacroreticular absorbent Amberlite XAD-2. Operation of the towed “fish”type sampler causes minimal interruption to a ship’s programme and allowsa large area to be surveyed. The second type is a self-powered in situ pumpwhich can be left unattended to extract large volumes of water at a fixed station.

Lovelock and co-workers [228, 229] determined methyl fluoride, methylchloride, methyl bromide, methyl iodide, and carbon tetrachloride in the At-lantic Ocean. This shows a global distribution of these compounds. Murrayand Riley [230, 231] confirmed the presence of carbon tetrachloride, and alsofound low concentrations of chloroform and tri- and tetrachloroethylene inAtlantic surface waters.

Kristiansen et al. [232] identified halogenated hydrocarbon byproducts inchlorinated seawater.

9.14Volatile Organic Compounds

9.14.1Head Space Analysis

There have been many applications of the technique [233–240]. The majoradvantage of the headspace method is simplicity in handling the materials. Atmost, only one chemical, the salt used in the salting-out procedure, needs to beadded, and in most cases the headspace gas can be injected directly into a gaschromatograph or carbon analyser. On the other hand, the concentration oforganic materials present is limited by the volume of sea water in the samplebottle. This is very much a batch process. Equilibration between the headspacegas and the solution can take a considerable time. This is not a problem whenthe salting-out material is added at sea, and the samples are then brought intothe laboratory for analysis some time later. When the salting-out is done inthe laboratory, equilibration can be hastened by recirculating the headspacegas through the solution. A system could be devised which would permit theaccumulation of volatiles from a large volume of water into a relatively smallheadspace, perhaps by recirculating both water and headspace gas througha bubbling and collection chamber, but much of the simplicity and freedomfrom possible contamination would be lost in the process.

Volatile organic materials can also be removed from solution by distillation,either at normal or at elevated pressures. While the amounts to be collectedin this fashion are small, if headspace samples are taken at elevated temper-atures and pressures, trace quantities of organics can be collected. It should

9.15 Chlorinated Dioxins 393

be emphasised, however, that whenever extreme conditions are employed tofree an organic fraction, that fraction is defined by the conditions of the sep-aration and cannot profitably be compared with fractions defined by differentsampling conditions. The use of elevated temperatures and pressures may alsoalter the compounds separated, limiting the amount of information that canbe extracted from the analyses.

9.14.2Stripping Methods

Juettner [242] has studied the application of stripping methods to the deter-mination of volatile organic compounds in seawater.


Kasthurikrishnan et al. [243] used membrane mass spectrometry to study theanalysis of volatile organic compounds in seawater at ppt concentrations.

9.15Chlorinated Dioxins

Hashimoto et al. collected chlorinated dioxins and furnaces at concentrationsof 1 ppq in seawater on a XAD-2 column [244].

9.16Nitrogen Compounds

Simply on the basis of the normal composition of marine organisms, we wouldexpect proteins and peptides to be normal constituents of the dissolved organiccarbon in seawater. While free amino acids might be expected as products ofenzymic hydrolysis of proteins, the rapid uptake of these compounds by bac-teria would lead us to expect that free amino acids would normally constitutea minor part of the dissolved organic pool. This is precisely what we do find; theconcentration of free amino acids seldom exceeds 150 µg/l in the open ocean.It would be expected that the concentration of combined amino acids would bemany times as great. There have been relatively few measurements of proteinsand peptides, and most of the measurements were obtained by measuring thefree amino acids before and after a hydrolysis step. Representative methodsof this type have been described [245–259]. Since these methods are basicallyfree amino acid methods, they will be discussed next in conjunction with thosemethods.


Free Amino Acids

Although the free amino acids are present only at very low concentrationsin oceanic waters, their importance in most biological systems has led to aninordinate amount of effort toward their determination in seawater. A sensi-tive, simple, and easily automated method of analysis, the colorimetric nin-hydrin reaction, has been known in biochemical research for many years. Inorder for the method to be useful in seawater, the amino acids had to beconcentrated. This concentration was usually achieved by some form of ionexchange [251]. An automated method not using a concentration step wasdeveloped by Coughenower and Curl [252]. While the method was used suc-cessfully in Lake Washington, its limit of detection (0.5 µmol/l) is just too greatfor most oceanic samples.


Various workers have discussed the application of this technique [164, 253–257], employing ninhydrin [218] and o-phthaldehyde as fluorescing agents[257].

Amino acid analysers based on ion exchange resins are available com-mercially. These achieve good separations of amino acid mixtures. Fluores-cent derivatives of separated amino acids constitute a very sensitive meansof detecting these compounds in seawater [256, 258]. Fluorescent derivativesthat have been studied include o-phthalaldehyde [259], dansyl [260], fluo-rescamine [261], and ninhydrin [261].

The amino acid analyser using fluorescamine as the detecting reagent hasbeen used to measure 250 picomoles of individual amino acids routinely [262],and dansyl derivatives have been detected fluorometrically at the 10–15 Mlevel [260]. Where the amounts of amino acid are high enough, the fluo-rescamine method, with no concentration step, can be recommended for itssimplicity. At lower concentrations, the dansyl method, with an extraction ofthe fluorescent derivatives into a non-polar solvent, should be more sensitiveand less subject to interferences. For proteins and peptides, the fluorescaminemethod seems to be the most sensitive available method.

In this connection, it is interesting to note that Gardner [263] isolated freeamino acids at the 20 nmol/l level from as little as 5 ml of sample by cationexchange, and measured concentrations on a sensitive amino acid analyserequipped with a fluorometric detector.

The classic work of Dawson and Pritchard [264] on the determinationof α-amino acids in seawater uses a standard amino acid analyser modifiedto incorporate a fluorometric detection system. In this method the seawatersamples are desalinated on cation exchange resins and concentrated prior toanalysis. The output of the fluorometer is fed through a potential divider andlow-pass filter to a comparison recorder.

9.16 Nitrogen Compounds 395

Dawson and Pritchard [264] point out that all procedures used for concen-trating organic components from seawater, however mild and uncontaminat-ing, are open to criticism simply because of the ignorance as to the nature ofthese components in seawater. It is, for instance, feasible that during the processof desalination on ion exchange resins under weakly acidic conditions, metalchelates are dissociated and thereby larger quantities of “free” components arereleased and analysed.

An example of a chromatogram obtained from a seawater sample and themole percentage of each amino acid in the sample is depicted in Fig. 9.2.

Mopper [265] has described developments in the reverse phase perfor-mance liquid chromatographic determination of amino acids in seawater. Hedescribes the development of a simple, highly sensitive procedure based on theconversion of dissolved free amino acids to highly fluorescent, moderately hy-drophobic isoindoles by a derivatisation reaction with excess o-phthalaldehydeand a thiol, directly in seawater. Reacted seawater samples were injected with-out further treatment into a reverse-phase high-performance liquid chromato-

Figure 9.2. Chromatogram of a seawater extract (20 ml) sample for amino acids collected at6 m in the Kiel Fjord. The concentrations of the individual acids were quantified as follows(in nmol/l): meto, 11; asp 34.4; thr, 23.2; ser, 88; glu, 36; gly, 100; ala, 56; vol, 16; ileu, 9.6; leu,12; galactosamine and amino sugars, 4; tyr, 6.8; phe, 7.2; B-ala, 20.8; α-amino-γ , 14.4; orn,44; lys, 12; hist, 7.2; arg, 8.6; cysSO2H, 4; cit, trace; tan, cys, trace; glucose-amine, trace; met,trace; urea, trace; phosphoserine, trace; OH-lys, trace. The total concentration of amino acidin the sample lies around 51 µg/l, assuming a mean molecular weight of 100. Source: [264]


graphic column, followed by a gradient elution. The eluted amino acid deriva-tives were detected fluorometrically. Detection limit for most amino acids was0.1–0.2 nmol per 500 µl injection. Problems of inadequacies with the methoditself, sample handling, and whether chromatographically determined con-centrations might be considered as biologically available concentrations inseawater, are discussed.

Chromatographic Methods

Several of the earlier methods used paper chromatography for the final sep-aration and determination, after some concentration step. Starikova and Ko-rzhikova [248] concentrated the free amino acids by taking the acidified so-lutions to dryness, then extracting the sea salts with 80% ethanol. Riley andSegar [246] dropped the detection limit to 0.1 µg/l, using TLC. Litchfield andPrescott [258] made dansyl derivatives, extracted these into an organic sol-vent, and then separated the amino acids by TLC. Ligand exchange was usedas a concentrating mechanism by Clark et al. [266] followed by TLC for finalseparation. The formation of 2,4-dinitro-1-fluorobenzene derivatives, followedby solvent extraction of these derivatives and circular TLC was suggested byPalmork [267].

Ligand exchange has been a favoured method for the concentration of aminoacids from solution because of its selectivity [268]. Ion exchange has often thenbeen used for the final separation of the acids [269–272].

Other methods of concentration have also been used to bring the lev-els of amino acids into a range suitable for ion exchange chromatography.Bohling [273, 274] and Garrasi and Degens [275] used evaporation or freezedrying, followed by extraction of dried salts with organic solvents. Tatsumotoet al. [276] used coprecipitation with ferric hydroxide as the method of con-centration prior to ion exchange chromatography.

Pocklington [277] has separated amino acids in seawater using freeze dry-ing. The first step in his concentration of free acids from seawater was thefreeze drying of the seawater samples. To reduce interferences in the later stepsof the procedure, the sea salts were packed into a chromatographic columnand washed with diethyl ether to remove non-polar compounds. The diethylether extract, particularly from surface water samples, quite often containedcoloured materials as well as other organics. If a series of solvents of graduatedpolarity were passed through a sea salt column, a fractionation by polarityshould be obtained. With the proper choice of solvents, a form of gradient elu-tion could be devised which would result in a continuous rather than a batchfractionation.

Gas Chromatography

Since the amino acids are not volatile, derivatives that are both volatile andeasily synthesised had to be found. A silyl derivative was used by Pockling-


ton [278] after the amino acids had been separated as a group and concentratedby extraction of the freeze-dried sea salts with an organic solvent, followed byevaporation of the solvent.

A still greater increase in sensitivity can often be achieved by using anelectron capture detector instead of flame ionisation, if the compound or itsderivative contains a halogen atom. Derivatives of this sort had been foundfor the amino acid analyses; one procedure yielded n-butyl N-trifluoroacetylesters [279]. The method has been adapted for seawater and a similar methodusing the methyl ester was devised by Gardner and Lee [280]. While sufficientsensitivity could be achieved with the use of these techniques there was far toomuch sample handling in the separation and derivative formation. With thedevelopment of easily synthesised fluorescent derivatives, liquid chromatog-raphy has largely supplanted gas chromatography for these analyses.

9.16.2Proteins and Peptides

The older methods for the measurement of protein in natural waters usuallydepended upon the presence of aromatic amino acids in the protein, andcalculated total protein on the basis of an average tyrosine, tryptophan, orphenylalanine content. A method representative of this type was the Folinreagent method published by Debeika et al. [281]. While these methods wereuseful in fresh water and in some coastal regions, they were not sensitiveenough for the lower concentrations to be found in oceanic waters.

A fluorescence method that would measure either free or combined aminoacids, depending upon the pH of the solution, was originally proposed byUdenfriend et al. [282] and adapted for seawater by North [283] and Packardand Dortch [284]. In this Fluran method, peptides normally yield maximumfluorescence by pH 7, while amino acids fluoresce best at pH 9. With the properchoiceofbuffers, thefluorescenceofpeptides andproteins canbedifferentiatedfrom that due to free amino acids.

An attempt to use the infrared spectrum of materials collected at the seasurface for a quantitative measure of composition has been made by Baieret al. [285]. They dipped a germanium crystal through the surface film, thenran an internal reflectance spectrum on the material clinging to the crystal.From the spectrum, they concluded that the bulk of the material present in thesurface film was there as glycoproteins and proteoglycans.

9.16.3Nucleic Acids

Pillai and Ganguly [286] have concentrated the nucleic acids from seawater byadsorption on homogeneously precipitated barium sulfate, then hydrolysedwith 0.02 M hydrochloric acid and analysed for the constituents.


Hicks and Riley [287] have described a method for determining the naturallevels of nucleic acids in lake and seawaters, which involves preconcentrationby adsorption onto a hydroxyapatite, elution of the nucleic acids, and thenphotometric determination of the ribose obtained from them by hydrolysis.

9.16.4Enzyme Activity

Enzyme activity has been sought in seawater; Strickland and Solorzano [288]looked for photomonoesterase activity, while Maeda and Taga [289] used a flu-orometric method for assaying deoxyribonuclease activity in both seawaterand sediment samples.

9.16.5Aliphatic and Aromatic Amines

Aliphatic amines have been determined by a number of methods. Batleyet al. [290] extracted the amines into chloroform as ion-association complexeswith chromate, then determined the chromium in the complex colorimetri-cally with diphenylcarbazide. The chromium might also be determined, withfewer steps, by atomic absorption. With the colorimetric method, the limit ofdetection of a commercial tertiary amine mixture was 15 ppb. The sensitivitywas extended to 0.2 ppb by extracting into organic solvent the complex formedby the amine and Eosin Yellow. The concentration of the complex was mea-sured fluorometrically. Gas chromatography, with the separations taking placeon a modified carbon black column, was used by Di Corcia and Samperi [291]to measure aliphatic amines.

Varney and Preston [292] discussed the measurement of trace aromaticamines in seawater using high-performance liquid chromatography.

Aniline, methyl aniline, 1-naphthylamine, and diphenylamine at trace lev-els were determined using this technique and electrochemical detection. Twoelectrochemical detectors (a thin-layer, dual glassy-carbon electrode cell anda dual porous electrode system) were compared. The electrochemical behaviorof the compounds was investigated using hydrodynamic and cyclic voltamme-try. Detection limits of 15 and 1.5 nmol/l were achieved using colourimetricand amperometric cells, respectively, when using an in-line preconcentrationstep.

Petty et al. [293] used flow injection sample processing with fluorescencedetection for the determination of total primary amines in sea water. Theeffects of carrier stream flow rate and dispersion tube length on sensitivityand sampling rates were studied. Relative selective responses of several aminoacids and other primary amines were determined using two dispersion tubelengths. Linear calibration curves were obtained over the ranges 0–10–6 M and


0–10–8 M glycine. Precisions of better than 2% at 10–6 M and a detection limitof 10–8 M glycine were obtained.

Yang et al. [294] determined nanomolar quantities of individual molecularweight amines (and organic acids) in sea water. Amines were diffused from thesample across a hydrophilic membrane to concentrate and separate them frominorganic salts and most other dissolved organic compounds. Methylamine,dimethylamine, and trimethylamine were all found in measurable amounts insea water.

Florence and Farmer [295] determined parts per million of aliphatic aminesin sea water using spectrophotometric procedures.

9.16.6Nitro-Compounds

The gas chromatographic determination of isomers of dinitrotoluene in sea-water has been described by Hashimoto and co-workers [296, 297]. These au-thors describe the complete separation of six dinitrotoluene isomers using gaschromatography with support-coated open tubular glass capillary columnsand electron-capture detection. The method was applied to the qualitativeand quantitative analyses of trace levels of isomers in seawater and the re-sults were found to be satisfactory, with no need for further clean-up proce-dures.

The separation was achieved on an Apiezon L grease SCOT glass capillarycolumn. Complete separation of six dinitrobenzene isomers took 8 min.

The method was used for routine monitoring of dinitrotoluene concentra-tions in seawater from Dokai Bay, Japan. Both 2,6- and 2,4-dinitrotoluene weredetected. Concentrations of 2,4-dinitrotoluene in surface water samples werehigher than those in bottom water samples in 8 out of 10 samples.

To detect nitro explosives in seawater Barshick et al. [298] investigateda method based on gas chromatography-ion trap mass spectrometry.

Complex matrixes typically cannot be analysed directly to obtain the se-lectivity and sensitivity required for most trace analysis applications. To cir-cumvent this problem, solid-phase micro extraction techniques were used topreconcentrate analytes selectively prior to gas chromatography/ion trap massspectrometry analysis.

The choice of solid-phase microextraction sorbent phase was shown to beimportant especially for the amino metabolities of trinitrotoluene and RDX,which were extracted better on polar phases. Although equilibration timeswere quite lengthy, on the order of 30 min or greater, a sampling time of only10 min was shown to be sufficient for achieving low part-per-billion (ppb)to part-per-trillion (ppt) detection limits for trinitrotoluene and the aminometabolities in real seawater samples. Solid-phase microextraction was idealfor rapid screening of explosives in seawater samples.


9.16.7Azarenes

Shinohara et al. [299] have described a procedure based on gas chromatog-raphy for the determination of traces of two, three, and five-ring azarenes inseawater. The procedure is based on the concentration of the compounds onAmberlite XAD-2 resin, separation by solvent partition [300], and determina-tion by gas chromatography–mass spectrometry with a selective ion monitor.Detection limits by the flame thermionic detector were 0.5–3.0 ng and thoseby gas chromatography–mass spectrometry were in the range 0.02–0.5 ng.The preferred solvent for elution from the resin was dichloromethane and therecoveries were mainly in the range 89–94%.

Diagram – See Attachment A

9.16.8Urea

Urea has been of interest to the biological oceanographer because of its role asan excretion product of protein metabolism, its function in osmoregulation,and its reported use as a nitrogen source for phytoplankton growth.

Emmet [301] developed a colorimetric method involving chlorination ofthe urea with hypochlorite, followed by condensation with phenol. The limitof detection for this method was 0.2 µg/l as nitrogen. The method was easilyadaptable to automatic analysis.

Most workers now use a colorimetric method based on the reaction ofurea with biacetylmonoxime [302, 303]. The method has been adapted forautomated analysis by De Marche et al. [304].

Urea has been determined in seawater, employing the urea diacetyl reactionand spectrophotometry [305].

9.16.9Hydroxylamine

Von Breymann et al. [306] have described a method for the determination ofhydroxylamine in seawater based on gas chromatography with electron capturedetection.

9.16.10Acrylamide

Brown and Rhead [307] improved the sensitivity of high-performance liquidchromatography to0.2 µg/l. Thisprocedure consists ofbromination, extractionof the α, β-dibromopropionamide with ethyl acetate, and quantification usinghigh-performance liquid chromatography with ultraviolet detection. Samples

9.17 Sulfur Compounds 401

tested included river, sea, and estuarine waters, sewage and china clay workseffluents, and potable waters. The levels of inorganic ultraviolet-absorbing im-purities found in water samples did not interfere in this procedure. The solventextraction procedure lowered interferences in all samples tested without re-moval of acrylamide or excessive use of solvents. The experimental yields ofα, β-dibromopropionamide encountered gave a mean of 70.13 ± 8.52% (95%confidence level) for acrylamide-spiked waters, over the concentration range0.2–8.0 µg/l of acrylamide monomer.

9.16.11Ethylene Diamine Tetracetic Acid and Nitriloacetic Acid

Differential pulse polarography has also been used to determine EDTA andnitriloacetic acid in synthetic sea water and phytoplankton media [308]. Cad-mium is used to convert a large fraction of either ligand to the reducible cad-mium complex. The presence of competing metal cations, including copper, isnot detrimental if the method of standard additions is used. The method wasused in correlating the concentration of complexed and uncomplexed speciesof copper with phytoplankton productivity, and production of extracellularmetal-binding organic compounds.

9.17Sulfur Compounds

9.17.1Alkyl Sulfides and Disulfides

Dimethyl sulfide (DM) is a major volatile organosulfur compound produced inseawater by certain groups of phytoplankton. It has been suggested that the fluxof dimethyl sulfide to the marine atmosphere could affect the cloud albedo andtherefore the global nuclei environment, via oxidation reactions to form cloudcondensation [312]. Due to the significant spatial and temporal variations inthe concentration of dimethyl sulfide in seawater [313–316], and hence the fluxto the atmosphere, there is a need to fully categorise dimethyl sulfide and itsprecursors in seawater in order to provide more accurate determinations ofthe global sulfur cycle, especially for impact on the global climate.

Dimethyl sulfide is derived primarily from the enzymatic hydrolysis ofdimethyl sulfoniopropionate (CH3)2S+CH2CH2COO–;DMSP), anosmoregula-tory compound produced by a wide variety of marine phytoplankton [313,317].Intracellular DMSP hydrolysis has been shown in phytoplankton [318], inmacro algae [319], and also in bacteria following uptake of DMSP from sea-water [320]. Reported seawater concentrations of dissolved dimethyl sulfide(< 0.1–90 nM) and DMSP (1–1000 nM) vary with increasing depth, spatiallyfrom coastal areas to the open ocean, and also temporally from winter tosummer [313–316].


Gas Chromatography

Dimethyl sulfide (DM) and dimethyl disulfide have been measured in seawaterand in the atmosphere by gas chromatography [309] and by GC–MS [310].Some variety of cryogenic trapping is often used.

Leck and Baagander [311] determined reduced sulfide compounds in sea-water by gas chromatography using a flame ionisation detector. Substancesdetermined include methyl mercaptan, dimethyl sulfide, hydrogen sulfide andcarbon disulfide. Detection limits range from 0.2 ng/l (carbon disulfide) to0.6 ng/l (methyl mercapton).

Smith et al. [321] determined dimethyl sulfide at sea using a purge andtrap GC–MS system. This method provides good accuracy and precision forthe determination of dimethyl sulfide and its precursor dimethyl sulfoniopro-pionate (DMSP) in seawater for trials-based equipment. By using deuteratedinternal standards of DM-d6 and DMSP-d6, the precision for replicate deter-minations was shown to be as high as 1.6% for dimethyl sulfide and 5.8% forDMSP when the internal standard concentration differed by up to an order ofmagnitude from the components being determined. The method for DM using“commercial off-the-shelf” equipment gave a detection limit of 0.03 nM andwas linear to > 100 nM. The most appropriate sample preparation methodol-ogy for storing the samples for up to 56 h during intensive sampling periodsincluded filtration, acidification, and refrigeration.

9.17.2Thiols

Shea and MacCrehan [322] and Duane and Stock [323] determined hydrophilicthiols in marine sediment pore waters using ion-pair chromatography coupledto electrochemical detection.

9.17.3Dimethyl Sulfoxide

Andreae [324, 325] has described a gas chromatographic method for the de-termination of nanogram quantities of dimethyl sulfoxide in natural waters,seawater, and phytoplankton culture waters. The method uses chemical reduc-tion with sodium borohydride to dimethyl sulfide, which is then determinedgas-chromatographically using a flame photometric detector.

9.17.4Thiabend Azole

Capitan et al. [326] determined down to 0.1 ng/l thiabend azole residues inseawater using solid phase spectrophotometry.

9.18 Chlorinated Insecticides 403

9.17.5Cysteine and Cystine

Cysteine and cystine has been determined in seawater by a method based oncathode stripping voltammetry of the copper complex [327].

9.17.6Miscellaneous

Turner et al. [328] determined nanogram quantities of dimethyl sulfide anddimethyl sulfoniopropionate in Antarctic waters.

Dimethyl sulfide and other volatile organic compounds have been deter-mined in amounts down to 0.1 pg/l in seawater in a method described byWatanabe et al. [329].

Adsorption on XAD-2 and XAD-4 resins followed by solvent desorption andhead space GS has been employed for the preconcentration and determinationof volatile organosulfur compounds in estuary and seawater [330].

Savchuk et al. [331] used GC with an open tubular column and a chemi-luminescence detector to determine sulfur-containing organic compounds inamounts down to 0.1 ppt in seawater.

9.18Chlorinated Insecticides


Work on the determination of chlorinated insecticides has been almost ex-clusively in the area of gas chromatography using different types of detectionsystems, although a limited amount of work has been carried out using liquidchromatography and thin-layer chromatography.

Wilson and co-workers [332, 333] have discussed the determination ofaldrin, chlordane, dieldrin, endrin, lindane, o, p and p, p′ isomers of DDTand its metabolites, mirex, and toxaphene in seawater and molluscs. The USenvironmental Protection Agency has also published methods for organochlo-rine pesticides in water and wastewater. The Food and Drug Administra-tion (USA) [334] has conducted a collaborative study of a method for mul-tiple organochlorine insecticides in fish. Earlier work by Wilson et al. [333,335] in 1968 indicated that organochlorine pesticides were not stable in sea-water.

Since petroleum ether was the solvent used in the earlier studies for extract-ing the DDT from seawater, Wilson and Forester [332] initiated further studiesto evaluate the extraction efficiencies of other solvent systems.


The recovery rates of o, p′-DDE in all tests were greater than 89% withpetroleum ether, or with 15% diethylether in hexane followed by hexane ormethylene dichloride, indicating no significant loss during analyses. The av-erage percentage recovery of p, o′-DDT extracted from seawater (salinity 16–21 ppt) up to 14 days after initiation of the experiment was in the range 76%to 95%. Immediately after the seawater was fortified with 3.0 ppb of DDT allsolvent systems removed 93% of the DDT.

Liquid scintillation counting of (14C) DDT has been used to study the pick-up and metabolism of DDT by freshwater algae and to determine DDT inseawater [336].

Wilson [337] showed that liquid–liquid extraction of estuarine water imme-diately after addition of DDT gave acceptable recovery levels with all solventsystems tested, but analyses carried out several days later gave only partialrecovery owing to adsorption of DDT on suspended matter.

Aspila et al. [338] reported the results of an interlaboratory quality controlstudy in five laboratories on the electron capture gas chromatographic deter-mination of ten chlorinated insecticides in standards and spiked and unspikedseawater samples (lindane, heptachlor, aldrin, δ-chlordane,α-chlordane, dield-rin, endrin, p, p′-DDT, methoxychlor, and mirex). The methods of analysesused by these workers were not discussed, although it is mentioned that themethods were quite similar to those described in the water quality BranchAnalytical Methods Manual [339]. Both hexane and benzene were used for theinitial extraction of the water samples.

Results obtained by the five laboratories were reasonably consistent, andlindane recovery varied between 113±31% and 83±25%. Heptachlor revealedpoor recovery, which confirmed its degradation [339–344]. The degradationproduct [340–342] is known to be 1-hydroxychlordene.


Samuelsen [345] has reported an HPLC method for the determination oftrichlorfos and dichlorvos in seawater. A methyl cyanide phosphate bufferwas used and detection was achieved by ultraviolet measurements at 205 nm.

9.19Polychlorobiphenyls

In actual practice, environmental samples which are contaminated with PCBare also highly likely to be contaminated with chlorinated insecticides. Manyreports have appeared discussing co-interference effects of chlorinated insecti-cides in the determination of PCB and vice versa, and much of the more recentpublished work takes account of this fact by dealing with the analysis of bothtypes of compounds. This work is discussed below.

9.19 Polychlorobiphenyls 405

9.19.1Gas Spectrofluorometry

The fluorescence enhancement of 3, 4, 4′-trichlorobiphenyl [361] and 3, 3′, 4,4′-tetrachlorobiphenyl [362] in the presence of POLE has been used for theirquantitation at levels as low as 4.9 and 1.4 ng/ml, respectively, in seawater.


PCBs have gas chromatographic retention times similar to the organochlorineinsecticides and therefore complicate the analysis when both are present ina sample. Several techniques have been described for the separation of PCBfrom organochlorine insecticides. A review of these methods has been pre-sented by Zitko and Choi [346]. These techniques are time-consuming and, ingeneral, semi-quantitative. In addition, differential adsorption or metabolismof the Aroclor isomers in marine biota prevent accurate analysis of the PCBs.The gas chromatographic determination of chlorinated insecticides togetherwith PCBs is difficult. Chlorinated insecticides and PCBs are extracted togetherin routine residue analysis, and the gas chromatographic retention times ofseveral PCB peaks are almost identical with those of a number of peaks ofchlorinated insecticides, notably of the DDT group. The PCB interference mayvary, because the PCB mixtures used have different chlorine contents, but it iscommon for PCBs to be very similar to many chlorinated insecticides and thecomplete separation of chlorinated insecticides from PCBs is not possible bygas chromatography alone [347–351].

Figure 9.3 illustrates the possibility of the interference of DDT type com-pounds in the presence of PCBs. In an early paper on the determination ofPCBs in water samples which also contain chlorinated insecticides, Ashlingand Jensen [352] pass the sample through a filter containing a mixture of Car-bowax 4000 monostearate on Chromosorb W. The adsorbed insecticides areeluted with light petroleum and then determined by gas chromatography ona glass column (160×0.2 cm) containing either 4% SF-96 or 8% QF-1 on Chro-mosorb W pretreated with hexamethyldisilazane, with nitrogen as carrier gas(30 ml/min) and a column temperature of 190 ◦C. When an electron capturedetector is used the sensitivity is 10 ng lindane per cubic meter with a samplesize of 300 litres. The recoveries of added insecticides range from 50 to 100%;for DDT the recovery is 80% and for PCB, 93–100%.

Musty and Nickless [353] used Amberlite XAD-4 for the extraction andrecoveryof chlorinated insecticidesandPCBs fromwater. In thismethodaglasscolumn (20×1 cm) was packed with 2 g XAD-4 (60–85 mesh), and 1 litre of tapwater (containing 1 part per 109 of insecticides) was passed through the columnat 8 ml/min. The column was dried by drawing a stream of air through, thenthe insecticides were eluted with 100 ml ethyl ether-hexane (1:9). The eluate


Figure 9.3. Interference of DDT-type compounds in the presence of PCBs. Gas chromatog-raphy: 5% QF-1 on Gas Chrom Q, 100–120 mesh. Solid time: 1 p, o′-DDT; 2 p, p′-DDD;3 o, p′-DDT; 4 o, p′-DDD; 5 p, p′-DDE; 6 p, p′-DDMU; 7 o, p′-DDE; 8 o, p′-DDMU. Brokenline = PCB Chiophen A 50. Source: [346]

was concentrated to 5 ml and was subjected to gas chromatography on a glasscolumn (1.7 m × 4 mm) packed with 1.5% OV-17 and 1.95% QG-1 on Gas-Chrom Q (100–120 mesh). The column was operated at 200 ◦C, with argon(10 ml/min) as carrier gas and a 63Ni electron capture detector (pulse mode).Recoveries of BHC isomers were 106–114%, of aldrin 61%, and of DDT isomersand polychlorinated biphenyls 76%.

Girenko et al. [354] noted that PCBs interfered with the electron capturegas chromatographic identification and determination of chlorinated insec-ticides in water and fish. Prior to gas chromatography, water samples andbiological material were extracted with n-hexane. A mixture of organochlo-rine insecticides was completely separated. Girenko et al. [354] noted that itwas difficult to analyse samples of seawater because they are severely pol-luted by various co-extractive substances, chiefly chlorinated biphenyls. Todetermine organochlorine insecticide residues by gas chromatography withan electron capture detector, the chlorinated biphenyls were eluted from thecolumn together with the insecticides. They produce inseparable peaks withequal retention time, thus interfering with the identification and quantitativedetermination of the organochlorine insecticides. The presence of chlorinatedbiphenyls is indicated by additional peaks on the chromatographs of the wa-ter samples and aquatic organic organisms. Some of the peaks coincide withthe peaks of the o, p′ and p, p′ isomers DDE, DDD, and DDT, and some ofconstituents are eluted after p, p′-DDT.

Elder [355] determined PCBs in Mediterranean coastal waters by adsorptionontoXAD-2resin followedbyelectroncapturegaschromatography.Theoverallaverage PCB concentration was 13 ng/l.

9.19 Polychlorobiphenyls 407

Amberlite XAD-2 resin is a suitable adsorbent for polychlorinated biphenyland chlorinated insecticides (DDT and metabolites, dieldrin) in seawater.These compounds can be suitably eluted from the resin prior to gas chro-matography [356, 358].

Picer and Picer [357] evaluated the application of XAD-2, XAD-4, and Tenaxmacroreticular resins for concentrations of chlorinated insecticides and poly-chlorinated biphenyls in seawater prior to analysis by electron capture gaschromatography. The solvents that were used eluted not only the chlorinatedhydrocarbons of interest but also other electron capture sensitive materials, sothat eluates had to be purified. The eluates from the Tenax column were com-bined and the non-polar phase was separated from the polar phase in a glassseparating funnel. Then the polar phase was extracted twice with n-pentane.The n-pentane extract was dried over anhydrous sodium sulfate, concentratedto 1 ml and cleaned on an alumina column using a modification of the methoddescribed by Holden and Marsden. The eluates were placed on a silica gel col-umn for the separation of PCBs from DDT, its metabolites, and dieldrin usinga procedure described by Snyder and Reinert [359] and Picer and Abel [360].

Picer and Picer [357] investigated the recovery from 10 litre samples of sea-water of 0.1–1.0 µg/l chlorinated pesticides (DDT, DDE, TDE, and Dieldrin),and 1–2 µg/l PCB (Aroclor 1254). The recovery of Mirex during these steps var-ied between 80% and 90%. Losses of the investigated chlorinated hydrocarbonsduring these steps were 10–30% for about 10 ng pesticides.

Petrick et al. [363] used HPLC to remove aliphatic compounds from estuarywaters prior to determining PCBs by gas chromatography.

Pedensen-Bjergaard et al. [364] compared three different methods (GC–ECD, GC–MS, GC–AED) for the determination of polychlorobiphenyls inhighly contaminated marine sediments.

Leoni [366] observed that in the extraction preconcentration of organochlo-rine insecticides and PCB’s from surface and coastal waters in the presence ofother pollutants such as oil, surface active substances, etc., the results obtainedwith an absorption column of Tenax-Celite are equivalent to those obtainedwith the continuous liquid–liquid extraction technique. For non-saline watersthat contain solids in suspension that absorb pesticides, it may be necessaryto filter the water before extraction with Tenax and then to extract the sus-pended solids separately. Analyses of river and estuarine sea waters, filteredbefore extraction, showed the effectiveness of Tenax, and the extracts obtainedfor pesticide analysis prove to be much less contaminated by interfering sub-stances than corresponding extracts obtained by the liquid–liquid technique.Leoni et al. [365] showed that for the extraction of organic micro pollutantssuch as pesticides and aromatic polycyclic hydrocarbons from waters, therecoveries of these substances from unpolluted waters (mineral and potablewaters) when added at the level of 1 µg/l averaged 90%.

Water samples were passed through the peristaltic pump into the absorp-tion column at a flow rate of about 3 litres per hour. When the absorption


on Tenax-Celite was completed the pesticides were eluted. For the analysisof naturally polluted water the solvent was evaporated, the residue dissolvedin light petroleum, and the solution purified by partitioning with acetoni-trile saturated with light petroleum [366, 367, 373]. The resulting solution wasevaporated just to dryness, the residue dissolved in 1 ml of n-hexane, andinsecticides and polychlorobiphenyls were separated into four fractions bydeactivated silica gel micro column chromatography (silica gel type Grace950, 60–200 mesh) [368]. The various eluates from the silica gel were thenanalysed by gas chromatography [369]. In order to evaluate the effective-ness of extraction from non-saline waters with the Tenax Celite column,the samples were also extracted simultaneously by the liquid–liquid tech-nique.

In the adsorption with Tenax alone satisfactory results were obtained, whilein the presence of mineral oil a considerable proportion of the organophos-phorus pesticides (particularly Malathion and Parathion-methyl) was not ad-sorbed and was recovered in the filtered water. This drawback can be overcomeby adding a layer of Celite 545 which, in order to prevent blocking of the col-umn, is mixed with silanised glass wool plugs. A number of analyses of surfaceand estuarine sea waters were carried out by this modified Tenax columnand simultaneously by the liquid–liquid extraction technique. To some of thesamples taken, standard mixtures of pesticides were also added, each at thelevel of 1 µg/l (i.e., in concentration from 13 to 500 times higher than thatusually found in the waters analysed). One recovery trial also specifically con-cerned polychlorobiphenyls. The results obtained in these tests show that thetwo extraction methods, when applied to surface waters that were not filteredbefore extraction, yielded very similar results for many insecticides, with theexception of compounds of the DDT series, for which discordant results werefrequently obtained.

Leoni et al. [365,370] conclude that the extraction of insecticide from watersby adsorption on Tenax yields results equivalent to those by the liquid–liquidprocedure when applied to waters that do not contain solid matter in suspen-sion. For waters that contain suspended solids that can adsorb some insecti-cides in considerable amounts, the results of the two methods are equivalentonly if the water has previously been filtered. In these instances, therefore, theanalysis will involve filtered water as well as the residue of filtration.

9.19.3Column Chromatography

Millar et al. [371] carried out experiments to study a method for the recoveryof 18 organochlorine pesticides and seven PCBs from water. Extractions withdichloromethane, and 15% dichloromethane in hexane, at pH 2, 7, and 10, andliquid–solid column chromatography using columns of Florasil or alumina,produced excellent results. An investigation was also made into the effects of

9.20 Organophosphorus Compounds 409

pH, temperature, and residual chlorine on the preservation of spiked samples,and recommendations were made for the most suitable storage conditions.

9.19.4Miscellaneous

Kristiansen et al. [232] identified halogenated hydrocarbon byproducts inchlorinated seawater used for drinking water. Phenol, cresols, and catecholswere present at low-ppb concentrations in San Diego Bay (CA, USA) [373].

Sullivan et al. [374] studied the loss of PCBs from seawater samples duringstorage.

Kelly et al. [372]hasdescribedasamplingapparatusconstructed tocollect 28litre samples of seawater designed to mimimise sample contamination derivedfrom the ships environment. Its utility in the study of polychlorobiphenyls,pentachlorophenols, and organochlorine pesticides was investigated.

Petrick et al. [375] extracted up to 2000 dm3 of Atlantic Ocean waters us-ing various solid resins. Down to 5 ng/dm were determined of chlorinatedbiphenyls, HCB, DDE, and polyaromatic hydrocarbons in samples taken atdepths down to 4000 metres.

9.20Organophosphorus Compounds

The soluble organic phosphorus compounds arising from natural sources havebeen examined by a number of techniques, including concentration by freeze-drying and separation by molecular size on Sephadex gels, by Minear [376] andMinear and Walanski [377]. The most familiar of these compounds, adenosinetriphospate, is normally considered to be found in the particulate fraction, andis measured by the well-known luciferin–lucifernase reaction [378]. An adap-tation of the method for estuarine waters and sediments was published by Wil-dish [379] and improvements on the usual method, including improvementsin sampling and sample handling, were discussed by Hofer-Siegrist [380].


Weber [381] has described a kinetic method for studying the degradation ofParathion in sea water. Weber observed two pathways whereby Parathion is hy-drolysed. The first reaction proceeds via dearylation with loss of p-nitrophenol:

(C2H5O)2–PS–OC6H4–NO2 + H2O

= (C2H5O)2–PS–OH + HOC6H4NO2


Additionally, they observed a second pathway, hydrolysis through dealkyla-tion leading to a secondary ester of phosphoric acid which still containsthe p-nitrophenyl moiety, i.e., de-ethyl Parathion (O-ethyl-O-p-nitrophenyl-monothiophosphoric acid):

(C2H5O)2–PS–OC6H4NO2 + H2O

= NO2–C6H4O–PS(OC2H5)OH + C2H5OH

Aliquots of ethanolic solution of Parathion were separated into decomposedinsecticides and decomposed insecticide products. Among the products freep-nitrophenol, chemically bound in acidic phosphorus compounds and innon-hydrolysed neutral phosphorus compounds, was detected in the same wayafter specification. Saponification after removal of ether without separation ofneutral and acidic compounds yielded total p-nitrophenyl equivalents.


Lores et al. [382] discussed the determination of the organophosphate insecti-cide fenthion in seawater. The method comprises a solvent extraction followedby silica gel clean-up procedure prior to determination by gas chromatographywith thermionic detection. Detection levels of 0.01 µg/l were achieved.

The insecticide fenitrothion (O,O-dimethyl-O-4-nitro-3-methylphenyl thio-phosphate) can be measured in sea water and sediments by gas chromatog-raphy, using a flame photometric detector to determine P and S [387]. Thedegradation products of the organophosphorus insecticides can be concen-trated from large water by collection on Amberlite XAD-4 resin for subsequentanalysis [383].

9.20.3Enzymatic Methods

The anticholinesterase nerve gases isopropyl methyl phosphonofluoridate(GB) and O-ethyl S-diisopropylaminoethylmethylphosphonothioate (VX) canbe measured in seawater by an enzymic technique [384].

Cambella and Antia [385] determined phosphonates in seawater by frac-tionation of the total phosphorus. The seawater sample was divided into twoaliquots. The first was analysed for total phosphorus by the nitrate oxida-tion method capable of breaking down phosphonates, phosphate esters, nu-cleotides, and polyphosphates. The second aliquot was added to a suspensionof bacterial (Escherichia coli) alkaline phosphatase enzyme, incubated for 2 hat 37 ◦C and subjected to hot acid hydrolysis for 1 h. The resultant hot acid–enzyme sample was assayed for molybdate reactive phosphate which was es-timated as the sum of enzyme hydrolysable phosphate and acid hydrolysable

9.21 Azine Herbicides 411

phosphate. Phosphate concentrationswere calculatedas thedifferencebetweentotal phosphorus and molybdate-reactive phosphorus (enzyme hydrolysablephosphate plus acid hydrolysable phosphate) thus exploiting the inert nature ofthe strong carbon–phosphorus bond. This bond was resistant to phosphataseaction and acid hydrolysis.


Ahern et al. [386] has observed that organophosphorous compounds are co-precipitated with cobalt and pyrrolidine dithiocarbonate, and suggest that thismight be a suitable means of preconcentrating the phosphorus fraction ofharbour water prior to analysis by X-ray fluorescence spectrometry.

9.21Azine Herbicides


Wu et al. [388] carried out measurements of the enrichment of Atrazine onthe micro surface water of an estuary. These authors used a micro surfacewater sampling technique with a 16 mesh stainless steel screen collecting bulksampled from the top 100–150 µm of the surface. Atrazine concentration inthe actual micro surface was estimated to vary in the range 150–8850 µg/l.

Abel et al. [389] determined simazine in estuary water by adsorption ona C18 SPE cartridge followed by determination by HR-GC using a nitrogen-phosphorus specific detector.


Durand and Barcelo [390] carried out a study of the interferences in the analysisof chlortriazines in seawater using GC–MS secondary ion monitor spectrom-etry and by GC using a nitrogen–phosphorus specific detector C39.

Electrospray mass spectrometry has been used to characterise triazine,phenylurea, and other herbicides in estuarine water [391].


A C18 disk has been used to remove chlorotriazine, atrazine, metabolites,organophosphorus compounds, phenylurea, and carbamate pesticides fromseawater prior to analysis by HPLC [392].


A study has been carried out on the determination of triazine and carba-mate pesticides and metabolities in seawater by HPLC with photodiode-arraydetection [393].

9.22Diuron, Irgalol, Chlorothalonil

Ferrer and Barcelo [394] used solid-phase extraction coupled with LC–MS todetermine diuron, irgalol, and chlorothalonil in seawater in amounts down to1 ng/l.

9.23Lipids

Marine lipids are important biological energy sources and have been used astracers in food studies [395–398]. Some lipids, however, are pollutants [399,400], and all lipids can potentially act as solvents, transporters, or sinks forpollutants [374, 399, 401, 402].

The rapidity with which the Iatroscan THIO MK II (Iatron Laboratories,Tokyo) TLCFID flame ionisation detector system provides synoptic lipid classdata from small samples suggested that it would be useful for screening seawa-ter samples prior to performing more detailed chromatographic analyses. Thelipid class concentrations obtained by TLCFID provide reference values for theconcentrations of the individual components obtained by other techniques.In particular, shipboard TLCFID analysis would help in deciding samplingstrategies for more detailed investigations. In addition, the TLCFID techniquealone could provide an overall picture of spatial or temporal variations in thedistribution of a complete range of lipid classes.

Standards used for the construction of calibration curves had alkyl chainlengths in the range C16 to C19. Table 9.2 gives the principal compounds usedas representatives of each class.

Lipids were separated on the thin-layer plate with solvents of increasingpolarity.

This technique was used by Delmas et al. [404] to separate lipid extracts inseawater into various classes. Lipid classes that have been eluted away from thepoint of application may be burnt off the rod in a partial scan, allowing thoselipids remaining near the origin to be developed into the place that has justbeen simultaneously scanned and reactivated. By analysis of complex mixturesof neutral lipids in this stepwise manner it is possible to be more selectiveabout lipid class separations as well as to be more confident about assigningidentities to peaks obtained from a seawater sample. In addition, this approachalso reduces the possibility of peak contamination by impurities which wouldnormally coelute with marine lipid classes (e.g., phthalate esters [403]).

Figure 9.4 shows the type of results attained by this procedure.

9.24 Sterols 413

Table 9.2. Seawater lipid classes and standards used for their identification and calibration

Class Abbreviation Standards and suppliers

Aliphatic hydrocarbons HC Nonadecane (Polyscience)Wax and sterol esters WE Hexadecyl palmitate (Analabs)3-Ketone (internal stand) KET Hexadecan-3-one (K & K Labs)Free fatty acid FTA Palmitic acid (Supelco)Triglyceride TG Tripalmitin (Supelco)Free alcohol ALC Hexadecan-1-ol (Polyscience)Free sterol ST Cholesterol (Supelco)Polar lipid PL Dihexadecanoyl lecithin (Supelco)


Figure 9.4. Shipboard analysis of particulate lipid classes on the same Chromarod. a Non-polar neutral lipids and internal standard, partial scan. b Remaining neutral lipids andpolar lipids, full scan, increased attenuation at X. Source: [404]

9.24Sterols

No class reaction has been proposed for the measurement of total sterols.Instead, various fractionation methods, usually derived from the biochemicallitreature, have been adapted to the concentrated materials collected fromseawater. Certain of the more important sterols, particularly those used in theevaluation of water quality, have been determined by the use of a compound-specific reaction, after concentration from solution. Thus Wann et al. [405,


407] measured coprostanol, a faecal sterol in seawater, after collection andseparation, by extraction using liquid–liquid partitioning or extraction ona column of Amberlite XAD-2 resin.

When solvent extraction was performed, the procedure used by Dutkaet al. [406] was followed with a slight modification: 2 ml of concentrated hy-drochloric acid and 5 ml of 20% (w/v) sodium chloride were added to each litreof water sample. The sample was extracted with vigorous mixing three timeswith 100 ml each of hexane for 30 min. The combined extract was washed withtwo 50 ml portions of acetonitrile (saturated with hexane) followed by two50 ml portions of 70% ethanol. The hexane was then brought to dryness ona rotary evaporator under reduced pressure. The sample was re-dissolved in100–200 µl of carbon disulfide, and 1–5 µl of the solution was used for gaschromatographic analysis.

14C-labelled cholesterol was used to test the recovery of 5–100 µg of faecalsterols from seawater (labelled coprostanol not being available). The radioac-tivity of the samples and eluates was measured by a two-channel liquid scintil-lation counter. Percentage recovery was calculated on the basis of the amountof labeled material recovered in the acetone eluant. The results indicate thatcolumn extraction efficiency is not adversely affected by the salinity of thewater samples, i.e., in the range 95–97%.

Most often the sterols have been collected by liquid–liquid extraction us-ing petroleum ether and ethyl acetate [408], chloroform and methanol [409],n-hexane [410, 411] or chloroform [412, 413]. After concentration, gas chro-matography was generally used for the final separation and determination, al-though thin-layer chromatography has also been employed. The extra sensitiv-ity of the electron capture detector could be used by reacting the concentratedsterols with bromomethyldimethylchlorosilane (BMDS) before separation andmeasurement [414].

An analysis performed by gas chromatography can usually also be per-formed by coupled gas chromatography–mass spectrometry, often with someincrease in sensitivity, and with considerably greater certainty in the identifi-cation of the compounds. This technique has been applied to seawater [415]and to marine sediments [416].

While most of the interest in sterols has been in the materials in solu-tion, Kanazawa and Teshima [417] have investigated the compounds presentin suspension. The suspended matter was fractionated by filtration througha graded series of filters, the sterols removed by extraction with an organic sol-vent, and the final separation and determination was made by flame ionisationgas chromatography.

Gas chromatography appears to give adequate separation and measurementof the various sterols to be found in the marine environment; where it is lessthan satisfactory is in the identification of the substances being measured. Withcompoundswhose structures canbesosimilar, onlygaschromatography–massspectrometry can be expected to provide reasonable identifications.

9.25 Chelators 415

9.25Chelators

While a great many compounds that have been or might be found in the marineenvironments have been accused of chelation, this section deals only with thenonspecific measurements of chelation, with what has been called “chelationcapacity”. In general, this capacity is measured by spiking the solution witha transition metal, preferably one that is easily measured, and then determiningeitherhowmuch is complexedorhowmuch is leftover.While theprincipleofallof the methods is the same, the details are different, and often quite ingenious.

The metals used are usually copper or cobalt. A good example of a relativelysimple approach is the paper by Hanck and Dillard [418]. They add an excessof CoII, which forms strong but labile complexes with the organic materialspresent. After complex formation, dilute hydrogen peroxide is added, to oxidisethe complex to the CoIII form, which is much less labile. The excess hydrogenperoxide is then destroyed and the unreacted CoII measured by differentialpulse polarography.

Another relatively simple approach is that of Stolzberg and Rosin [419].The sample is spiked with an excess of copper, then passed through a Chelex100 column. The column retains the free copper ion, but passes the copperassociated with strong ligands. The chelated copper eluted from the column ismeasured by plasma emission spectrometry.

A kind of “standard additions” approach can also be used for the mea-surement of apparent complexing capacity. In this technique, labile copperis measured by differential pulse anodic stripping voltammetry after each ofa number of spikes of ionic copper have been added to the sample [420].

In terms of solution chemistry, “apparent” capacities derived from the ex-tremely dilute and diverse natural samples, determined at natural pH values,are very crude. To get results that are more exact, or tell more about the possiblenature of the complexing materials, it is necessary to concentrate, and some-times to separate out, the organic chelators. This approach has been used byBuffle et al. [421]. Working with river water, they were able to concentrate andclean up their samples to the point where they were able to treat the chelatingmaterial almost as reagents, and to determine mean molecular weights for theligands, stability constants for the complexes and the pH dependency of thestability of the complexes. This work would be considerably more difficult inseawater, but if the same organic materials were to be found in both fresh andseawater, it might be possible to determine the various values for fresh waterand then to determine the effect of ionic strength on these quantities.

When the chelators are actually known, as in the case of industrial materialsinjected into the environment, it is possible to derive much more informationfrom the analyses. Thus high pressure liquid chromatography has been usedto separate the copper chelates of EDTA, NTA, EGTA, and CDTA with the finalmeasurement of copper being made by atomic absorption [422, 423].


A spectrophotometric method based on the light absorption of the colouredCoIII complexes has been used to determine EDTA and NTA in fresh wa-ter [424]. In these few cases, actual well-defined compounds were present atconcentrations high enough so that the individual compound could be isolated,identified, and measured. This is seldom the case for the chelators in seawater:we are usually measuring an attribute, not a compound, with little idea of theactual identity of the compounds.

9.26Humic Materials and Plant Pigments

Humic materials cover three main classes of compounds which are discussedbelow under separate headings. They include:

1. Fluorescent yellow materials known collectively as “Gelbstoff” (yellow ma-terial)

2. Lignins and lignin sulfonates3. Humic and fulvic acids

While in a strict sense only the last class of compounds should fall into thecategory of humic materials, there has been much dispute concerning theorigin of the marine humic materials; both Gelbstoff and the lignins have beenmentioned as probable starting compounds for the marine humic and fulvicacids.

Chlorophyll is also discussed in this section.

(a) Gelbstoff

Under this heading are discussed both the naturally occurring fluorescentmaterial and the yellow substances which give coastal waters their generallygreenish colour. It is usually considered that these two categories are the same,or at least overlap almost entirely.

An in situ technique for measuring fluorescence in the ocean has been de-veloped by Egan [425]. His sensors, set to measure separately both chlorophylland Gelbstoff fluorescence, can be lowered to 600 m and operate unattended.

Gel filtration with Sephadex was used by Ghassemi and Christman [426] tomake separations by molecular size on water samples concentrated by vacuumevaporation. Fluorescence was also used as one method for following thefractionation. Molecular size was also used by Gjessing [427] but with pressuredialysis as the method of separation. A similar method of concentration andseparation was used by Brown [428] to follow the dispersion of these materialsas fresh and salt water mixed in the Baltic Sea.

A method of concentration by adsorption on aluminium oxide has beenproposed. It achieved almost complete adsorption.

9.26 Humic Materials and Plant Pigments 417

The argument has been made by several investigators (see the review byWangersky [429]) that the humic materials in seawater are not derived from ter-restrial sources, but result from light-induced condensation reactions of phe-nolic material and proteins released by fixed algae in the coastal regions. Thesematerials can be collected on nylon-packed chromatographic columns [430].Methods of collection and fractionation, as well as studies on the chemical andphysical properties of these compounds, were discussed by these authors [431].These compounds seem to be relatively easy to collect, to fractionate, and tofollow through the various chemical procedures.

(b) Lignins

Lignins and lignin sulfonates are considered to be important tracers of man’sactivities. Themajor sourceof these compounds is thepulpandpaper industry;lignin is not a marine product, and any large accumulation of these compoundssuggests a local dominance of terrestrial materials.

Pocklington and Hardstaff [432] react sediment samples with 1,3,5-tri-hydroxybenzene in alcoholic hydrochloric acid to produce a colour in theparticulate lignins, facilitating their identification under the microscope. Sam-ples high in lignins can then be subjected to the normal methods of analysis.This is an excellent screening technique (semi-quantitative).

A concentration step is often used in order to bring the sample within theconcentration range of the method. Thus Stoeber and Eberle [433] extractedthe lignosulfonic acids with trioctylamine in chloroform before final determi-nation, and Revina and Kriul’kov [434] employed spectrophotometric meth-ods. Extraction is also used to remove interferences in spectrophotometricmethods [435].

Direct spectrophotometric methods have been proposed for both particu-late and dissolved lignosulfonic materials. Kloster [436] used the Folin Ciocal-teumethod, which actually measures hydroxylated aromatic compounds. A ge-neral review of spectrophotometric methods was published by Bilikova [437].

Certain derivatives of the lignin sulfonic acids can be determined directlyin water. The nitroso derivatives, which are easily formed in solution, can bedetermined by differential pulse polarography [438]. Vanillin can be formed byalkaline hydrolysis [439] or alkaline nitrobenzene oxidation [440], extractedinto an organic solvent and determined by gas chromatography.

As a general rule, fluorometric methods are considerably more sensitivethan spectrophotometric methods, although standardisation is more difficult.Direct fluorometric procedures for lignin and lignin sulfonates have beendescribed [441–443].

(c) Humic and Fulvic Acids

The designation of certain classes of organic materials as humic and ful-vic acids unfortunately implies a certainty and regularity of structure which


is far from true. These terms, derived from soil science, indicate only theproducts resulting from a particular sequence of fractionation; humic acidsare those compounds extracted from soils by alkaline solutions and pre-cipitated upon acidification, while fulvic acids are those extracted by alka-line solutions but left in solution on subsequent acidification [444]. Whilethese terms were originally specific to soil science, and to materials iso-lated from soils, the usage was gradually extended to marine sediments andthen by further extension to materials isolated from solution by similar tech-niques.

Evidence is accumulating which shows that materials isolated from themarine environment are quite unlike those isolated from soils, at least in13C/12C ratios and probably in structure.

The methods for isolating humic and fulvic acids from marine samples canbe found in King [445], Rashid and King [446], and Pierce and Felbeck [447].These techniques, which derive from soil chemistry, may be used when thesample is a marine sediment. However, when it is the material in solution thatis being separated, the application of the methods is not so straightforward.Several workers have used macroreticular resins, usually XAD-2, to collecthigh molecular weight materials from seawater [448–450]. It has been demon-strated that this resin will, in fact, collect the humic and fulvic acids separatedfrom soils by the usual methods [450]. It has therefore been assumed that thematerials collected by this resin from seawater are the humic and fulvic acids,although the characteristics of the fractions so collected are different fromthose collected from terrestrial soils and marine sediments [448, 451].

Applications of NMR, ESR, thermal analysis, spectrophotometry, gas chro-matography, and GC–MS to humic and fulvic acid analysis have been reviewedby Schnitzer [452].

Salfeld [453] has shown that qualitative information on the nature of thecompounds present can be found by means of derivative spectrophotometry.The concentration of fulvic acids in natural waters can be calculated from theultraviolet absorptionat400and340 nm,after removalof thehumicacids [454].If a somewhat greater degree of uncertainty as to the actual composition of thematerials is acceptable, sample preparation can almost be eliminated, and thesensitivity of the method increased, by using fluorescence rather than absorp-tion as the measure of the humic materials present [441,443]. However, it mustalways be remembered that in this method of analysis we are equating humicmaterials with all of the soluble compounds fluorescing in the proper region;the correspondence is not necessarily exact. In addition, the fluorescence ofhumic and fulvic acids is rapidly decreased as sea surface sunlight intensities.The fluorescence may therefore behave as a highly non-conservative indicatorfor these materials in surface waters.

Many of the methods used for lignins can also be applied to humic acids.Thus the use of pulse polarography after the formation of nitroso derivativesis possible with humic acids as well as lignins [433, 443]. If the fulvic acids are


removed on activated charcoal, then eluted with acid, they can be determinedby ac polarography [455].

Separation of the humic acids into molecular size classifications by gel per-meation chromatography has been used largely to determine the distributionof sizes within this group of compounds [456]. The information coming fromsuch fractionation cannot be taken at face value, however, since the elution ofthe humic materials is strongly influenced by the functional groups present,the pH, and the ionic strength. Such fractionation can profitably be usedas a first step in the GC–MS determination of the compounds present. Forthose compounds too large or too complex to be determined directly by gaschromatography–mass spectrometry, oxidation [457] or hydrolysis [458] canbe useful preliminary treatments.

For a rough estimation of the relative amounts of humic and fulvic acids,the fluorometric method has much to recommend it, not the least being itssimplicity. Also, the ability of the method to measure short-range variabilitymight be quite valuable. It is not possible to be certain of the value of any moredetailed or more exact information until a great deal more is known about thenature of the compounds in each of these fractions.

(d) Chlorophyll Spectrometry

While the spectrophotometric determination of plant pigments, particularlythe chlorophylls, has become a routine procedure [459,460], refinements aim-ing at either greater accuracy (usually by the separation and determinationof the separate pigments) or greater sensitivity continue to appear in the lit-erature. A review of the various methods, both for total pigments and foreach separate pigment, was written by Rai [461]. A method using separationof pigments by TLC and determination by densitometry was described byMessiha-Hanna [462], while Quirry et al. [463] discussed a laser absorptionspectrometry method with a sensitivity a hundred times that of the usualspectrophotometric method.

Bjarnborg [464] has studied shipboard methods of continuous recording ofin vivo chlorophyll fluorescence and extraction-based chlorophyll determina-tions in Swedish archipelago waters.

Boto and Bunt [465] used thin-layer chromatography for the preliminaryseparation of chlorophylls and phaeophytins from seawater, and combined thiswith selective excitation fluorometry for the determination of the separatedchlorophylls a, b, and c, and their corresponding phaeophytin components. Anadvantage of the latter technique is that appropriate selection of excitation andemission wavelengths reduces the overlap among the emission spectra of thevarious pigments to a greater extent than is possible with broadband excitationand the use of relatively broadband filters for emission.

The fluorescence studies were performed on 90% acetone solutions with anAminco-Bowman spectrofluorometer (Model J4-8203G).


Figure 9.5 shows an excitation and emission spectrum for chlorophylls a,b, and c, and their respective phaeophytins. The excitation spectra were ob-tained by holding the emission wavelength at the emission maximum andslowly scanning the excitation wavelength over the required range. The emis-sion spectra shown have not been corrected for changes in photomultiplierresponse with wavelength. These spectra show that with suitable choice of ex-citation wavelengths, good selectivity can be achieved. Assuming that a givenmixture contains all three chlorophylls and their phaeophytins, a set of simul-taneous equations can he derived for the emission responses at the usual emis-sion maxima of each pigment when the mixture is excited at various chosenwavelengths. After acidification of the mixture to phaeophytins alone, furtherinformation can be obtained as indicated below (solutions of the phaeophytinswere prepared from the corresponding chlorophylls by adding 2 drops of 0.1 Mhydrochloric acid per 100 ml solution).

Although the equations are complex, in most cases the minor terms aresuch that many can he neglected. Further simplification is often possible whenone considers that, in most offshore seawater samples, the only primary pig-ments found are chlorophylls a and c, along with phaeophytin a. Chlorophyll band phaeophytins b and c are either absent or present in minor amounts.Phaeophorbides are also commonly found. However, these would probablyhave very similar fluorometric properties to the phaeophytins, and hencecould not be separately determined by any fluorometric method.

Taking into account the most likely pigment compositions of the phyto-plankton in most seawater samples, a simplified set of equations was obtained.As the quantum efficiency of chlorophyll c appears to be very high, its emis-sion spectrum is very little affected by interference from the emissions of otherpigments.

Therefore, its concentration can be quite accurately calculated and this valuethen used to give accurate estimates of the others.

Murray et al. [466] carried out an intercomparison of the determination ofchlorophyll in marine waters by methods based on HPLC, spectrophotometry,and fluorometry. Good agreement was obtained between these methods forchlorophyll a and chlorophyll b.

High-Performance Liquid Chromatography (HPLC)

Abayashi and Riley [467] used HPLC to determine chlorophylls, and theirdegradation products, and carotenoids in phytoplankton and marine particu-late matter.

Pigment extraction is carried out with acetone and methanol. After evap-oration of the combined extracts under reduced pressure, the pigments areseparated on a Partisil-10 stationary phase with a mobile phase consisting oflight petroleum (bp 60–80 ◦C), acetone, dimethyl sulfoxide and diethylamide(75:23.25:1.5:0.25 by volume). When chlorophyll c is present, a further develop-


Figure9.5. Excitation and emission spectra. (a and b) Chlorophyll a (- - -) and phaeophytin a(—-); concentration of both 0.134 µg/ml. (c and d) Chlorophyll b (- - -) and phaeophytin b(—-); concentration of both 0.172 µg/ml. (e and f) Chlorophyll c (- - -) and phaeophytin c(—-); concentration of both 0.042 µg/ml. Note that the phaeophytin c curves are shown ona tenfold higher sensitivity scale to the others. All solutions are in 90% acetone. Source: [465]


ment is performed with a similar, but more polar, solvent mixture. Detection iscarried out spectrophotometrically at 440 nm. The method has a sensitivity forthe chlorophylls of approximately 80 ng, and for the carotene of approximately5 ng. The coefficient of variation of the chromatographic stage of the procedurelies in the range 0.6–1.8%.

Abayashi and Riley [467] compared results obtained by the HPLC methodwith those obtained by a reflectometric thin-layer chromatographic methodand the SCOR/UNESCO polychromatic procedure. The results obtained fromthe latter were evaluated by the SCOR/UNESCO equations. The carotenoidswere determined collectively from the absorption of the 90% acetone extract at480 nm by means of the equations of Strickland and Parsons [459]. The resultsof these comparative studies show that there is satisfactory agreement forall pigments between the two chromatographic methods. However, althoughthe results for chlorophyll a by the polychromatic method were in reasonableaccord with those derived chromatographically, many of those for the otherchlorophylls were highly discrepant. Obviously, the polychromatic method,in particular, is unsatisfactory with respect to the interference of chlorophylldegradation products, as these are nearly always present in environmentalsamples.

Thin Layer Chromatography

Garside and Riley [468] have used thin-layer chromatography to achieve a pre-liminary separation of chlorophylls on solvent extracts of water and algae priorto a final determination by spectrophotometry or fluorometry. Garside and Ri-ley [468] filtered sea water samples (0.5–5 litres) through Whatman GF/C glassfiber coated with a layer, 1–2 mm thick, of light magnesium carbonate. Thisretains the smallest particles of organic matter and it is easy to extract thepigment from it. The filter is extracted with acetone (3–5 ml) and then withmethanol (10 ml) using ultrasonic vibration. The solution is passed throughanhydrous sodium sulfate to remove water and then evaporated in vacuum atless than 50 ◦C. The residue is dissolved in ethyl ether-dimethylamine (99:1,1–2 ml), and this is applied as a spot to a plate coated with silica gel PF254. Thechromatogram is developed with light petroleum (bp 60–80 ◦C, ethyl acetate-dimethylamine (55:32:13))until the centreof the chlorophylla spothas traveledabout 10 cm from the origin. The solvent is allowed to evaporate and the platescanned by reflection of the light passing through an Ilford 601 filter (603 nm).The integration reading for each peak is measured and the Rf values notedrelative to chlorophyll a. Xanthophylls are identified by scraping off the spotsand measuring the absorption spectrum of an extract of the scrapings. Chloro-phyll c remains at the origin and can be developed in light before chlorophyllsa, b, and c, carotene, xanthophylls, and certain degradation products can bedetermined. The analysis takes 1 h, the sensitivity for chlorophyll a is about0.12 µg, and the precision for most pigments is ±5%or better at the 0.5 µg level.

9.27 Vitamins 423

Karabashev [469] devised a method to estimate the influence of dissolvedoxygen on remote sensing measurements of chlorophyll a in seawater. Therelationship between fluorescence and absorption was studied in this investi-gation.

9.27Vitamins

Only one analytical method has been widely applied to the measurement ofvitamins in seawater. The method, bioassay, is not really within the realmof the analytical chemist, since it requires the maintenance of cultures of testorganisms. The tests also usually require a minimum of four days before resultsare available.

The older methods for vitamin B12 used the optical density of the culturemedium as a measure of growth rate of the assay organism [470, 471]. Thesensitivity of the method could be increased somewhat by the following 14Cuptake as a measure of growth [472, 473]. These methods are sensitive to0.1 ng, well below the amounts of the vitamin which could be measured by anyavailable chemical technique.

Thiamine has also been measured by bioassay, a marine yeast being used asthe assay organism [474,475]. Marine bacteria [476,477], marine yeasts [475],and dinoflagellates [478] have been used for the assay of biotin.

There are many uncertainties in the use of bioassay methods, not the leastof these being the genetic stability of the assay organisms. Stock culturescannot be treated like chemical reagents, to be put back on the shelf andforgotten between analytical runs. This is an area of analysis best left to themicrobiologists; if the information is absolutely necessary, the maintenance ofcultures and the actual assay should not be left solely in the hands analyticalchemists.

9.28Cobalamin

Sharama et al. [479] compared results obtained in the determination of cobal-amins in ocean waters by radioisotope dilution and bioassay techniques. Theseworkers showed that the isotopic methods measured both biologically activeand inactive cobalamins indiscriminately when porcine factor was used as theB12-specific binder.

9.29Pectenotoxins

Suzuki et al. [480] developed a solid-phase extraction LC–MS method fordetermining pectenotoxin-2 and pectenotoxin-6 in seawater. These are toxinsproduced from toxic phytoplankton.


Table 9.3. Preconcentration of organics in seawater

Class of organic Preconcentration method Analytical finish LD Reference

Hydrocarbon Extraction with organic µg/l [482–488]solvents [489–492]CH2Cl2 [493–500]CHCl3 [501–503]CCl4 [84]Cl3FC [504, 505]C6H6 [506]C5H12 [507]Petroleum etherMeOH–C6H6EtOOCCH3-Petroleumether

[508]

Hydrocarbon Extraction with Cl3CF3,then adsoption on SiO2

Thermistor 5 [26]

Hydrocarbon Pentane, hexane, CH2Cl2 1–2 µgabsolute

Alcohols,ketones,aldehydes

Gas stripping, thenadsorption on SiO2

GC–MS 1 µg/l [509]

Amino acids Adsorption on zirconiumphosphate

– – [510]

Amino acids Adsorption on ferric oxide Io exchangechromatography(IEC)

3 µg/l [511, 512]

Amino acids Solvent extraction of2,4-dinitro phenol

AAS and TLC – [267]

Amino acids Evaporation to drynessand ethanol extraction

Paperchromatography

FAA16.3 µg/lCAA6 µg/l

[166] [514]

Amino acids Ligand exchange on IEC FAA [268, 515,Cu–Chelex 21 µg/l 164, 270]

CAA185 µg/l

[516]

Amino acids Evaporated to dryness,desalted on Cu–Chelex

GC CAA129 µg/l

[516]

Amino acids Extraction of dried saltswith acid, 80% ethanol de-salted

Paperchromatography

FAA2.3 µg/lCAA5.8 µg/l

[248]

Amino acids Extraction of dried saltswith acid

IEC FAA6 µg/l

[273]

Amino acids Extraction of dried saltswith acid

TLC FAA4.5 µg/lCAA2.1 µg/l

[246]

9.29 Pectenotoxins 425



Amino acids Freeze drying,ether–ethanol extraction

GC FAA6 µg/l

[253]

Amino acids Extraction with ethylacetate,Cu–Chelexmethod

IEC FAA1.8 µg/l

[272, 517]

Amino acids Desalt with Dowex 50WX8at pH 3–4 lyophilisation

IEC FAA4.5 µg/lCAA3.5 µg/l

[275, 518]

Amino acids Ligand exchange onCu–Chelex

IEC FAA1 µg/lCAA10 µg/l

[249, 519]

Amino acids SemiautomatedCu–Chelex method

IEC FAA0.5 µg/l10 µg/l

[520, 521]

Humic acids Adsorption on macroretic-ular resin XAD-2

– – [448–450]

Organochlorineand organo-phosphorusinsecticides

Reverse osmosis [270, 523]

Ethers, glycols,amines, nitriles,hydrocarbons,chlorinatedhydrocarbons

Reverse osmosis [522, 523]

Chlorinatedinsecticides

Adsorption on ionexchange resins, AmberliteXAD-21

[524]

Chlorinatedinsecticides

Adsorption on silicic acidCelite, desorption withpetroleum ether

[525]

Chlorinatedinsecticides andpolychlorobi-phenols

Adsorption onTenax–Celite

GC 1 µg/l [370, 526]

Organophospho-rus insecticides

Adsorption on Sephadexgel

[376, 377]

Miscellaneous(phenols, car-boxylic acids, sur-factants, carbo-hydrates, aminoacids, proteins,insecticides,humic acids)

Adsorption on macroretic-ular resins, XAD1

2–5 µgabsolute

[527, 528]




Miscellaneous Adsorption on Sep-Pak C18cartridges

– – [529]

Polyaromatichydrocarbons,phenols

Extraction with methylenedichloride

GC–MS, HPLC – [530]

Aliphatichydrocarbons

Adsorption on Tenax GC ora Chromasorb

GC µg/l [13, 15, 23]

Aliphatichydrocarbons

Head space analysis GC–MS [531–534]

Organotincompounds

Adsorption on cationexchange resin

Polarography [535]

Organotincompounds

Adsorption on XAD-2resin, desorbed withhexane–methanol

– – [536]

Organomercurycompounds

Extraction with dithizonesolution

Gold trap coldvapour AAS

– [537]


9.30Flavins

A method has been reported [481] for the determination of flavins in seawater.The method is based on solid-phase extraction. With ion-pair HPLC using flu-orescence detection, concentrations in the picomolar range can be measured.

9.31Microcystine

Metcalf et al. [543] used enzyme linked immunoassay (ELISA) to measuremicrocystine in marine water.

9.32Preconcentration of Organics

Applications of preconcentration to µg/l levels are reviewed in Table 9.3.

References

1. Blumer M, Blokker PC, Cowell EB, Duckworth EF (1972) In: Goldberg ED (ed) A guideto marine pollution. Gordon and Breach Science Publishers, New York, pp 19–40

2. Farrington JW (1973) Analytical techniques for the determination of petroleum con-tamination in marine organisms. AD Report No.766792/6 US National Technical In-formation Service, Springfield, VA

References 427

3. Andelman JB, Snodgrass JE (1974) Crit Rev Env Contr 4:694. Institute of Petroleum (1974) Marine pollution by oil: characterization of pollutants,

sampling, analysis, and interpretation. Applied Science Publications5. McGlynn JA (1974) A review of techniques for the characterization and identifica-

tion of oil spillages. In: Examination of waters, evaluation of methods for selectedcharacteristics. Resource council technical paper no. 8, pp 85–89

6. Whitham BT, Duckworth DF, Harvey AAB, Jeffrey PO, Perry SG (1974) Characteriza-tion of pollutants, sampling, analysis, and interpretation. In: Whitham BT (ed) Marinepollution by oil. Applied Science Publishers, Essex

7. Erskine RL, Whitehead EV (1975) Iranian Journal of Science and Technology 3:2218. IOC (1975) Report of the ICG for GIPME Task Team for the Evaluation of Recommen-

dation No. IV of the Joint IOC/WMO Task Team II on the Marine Pollution MonitoringProject UNESCO Paris

9. Moulder DS, Varley A (1975) A bibliography on marine and estuarine oil pollution,supplement 1. Marine Biology Association

10. National Academy of Sciences, USA (1975) Petroleum in the marine environment.A workshop on inputs, fates and the effects of petroleum in the marine environment.Airlie, 1973

11. Fam S, Stenstrau MK, Silverman G (1987) J Environ Eng 113:103212. Wade TL, Quinn TG (1975) Mar Pollut Bull 6:5413. May WE, Chester SN, Cram SP (1975) J Chromatogr Sci 13:53514. Drodz J, Vadokova Z, Novak J (1986) J Chromatogr 354:4715. Swinnerton JW, Linnenbom VJ (1967) Science 156:111916. Desbaumes E, Imhoff C (1972) Water Res 6:88517. Polak J, Lu B (1974) Anal Chim Acta 63:23118. Schonmann M, Kern H (1981) Varian Instrument News 15:619. Brooks JM, Sackett WM (1973) J Geophys Res 78:524820. Atkinson LP, Richard FA (1967) Deep-Sea Res 14:67321. Perras JC (1973) Portable gas chromatographic technique to measure dissolved hy-

drocarbons in seawater. US NTIS AD report no 786583/5GA22. Swinnerton JW, Linnenbom VJ (1967) Science 156:112023. Bellar TA, Lichtenberg JJ (1975) The determination of volatile organic compounds

at the µg/l level in water by gas chromatography. US National Technical InformationService report no. 237973/3GA

24. Wasik SB (1974) J Chromatogr Sci 12:84525. Ahmadyian M, Brown CV (1976) Anal Chem 48:125726. Zsolnay A, Kiel MJ (1974) J Chromatogr A 90:7927. Booksh KS, Muroski AR, Myrick ML (1996) Anal Chem 68:353928. Krstulovic AM, Rosie DM, Brown PR (1976) Anal Chem 48:138329. Hiltabrand RT (1978) Mar Pollut Bull 9:1930. Payne JR (1976) Science 191:94531. Isaaq HJ, Andrews AW, Janini GN, Barr EW (1979) J Chromatogr 2:31932. Zitko V, Carson WV (1970) Tech Report Fish Res Bd Can No. 21733. Michalik PA, Gordon DC (1971) Tech Report Fish Res Bd Can No. 28434. Levy EM (1971) Water Res 5:72335. Levy EM (1972) Water Air Soil Poll 1:14436. Levy EM, Walton A (1973) Journal of the Fisheries Research Board Canada 30:26137. Van Duuren BL (1960) Anal Chem 32:143638. Keizer PD, Gordon DC Jr (1973) Journal of the Fisheries Research Board of Canada,

30:1039


39. Cretney WJ, Wong CS (1974) Special publication no. 409. US National Bureau ofStandards, p 175

40. Hellman H, Sehle H (1973) Fresen Z Anal Chem 265:24541. Hellman H (1975) Fresen Z Anal Chem 275:10942. Schwarz EP, Wasik SP (1976) Anal Chem 48:52443. Freegarde M, Hatchard CG, Parker GA (1971) Laboratory Practice 20:3544. Thruston AD, Knight HW (1971) Environ Sci Technol 5:6445. Lloyd JBF (1971) J Forensic Sci Soc 11:8346. Lloyd JBF (1971) J Forensic Sci Soc 11:15347. Lloyd JBF (1971) J Forensic Sci Soc 11:23548. Parker CA, Barnes WJ (1960) Analyst 85:349. Maher WA (1983) Bull Environ Contam Tox 30:41350. Red’Kin YR, Voitenko AM, Teplyakov PA (1973) Okeanologiya 13:90851. Liu J, Xie Z, Wang X, Wang R (1993) Fenxie Shiyanshi 12:2852. Kawahara FK (1969) Environ Sci Technol 3:15053. Kawahara FK, Ballinger DG (1970) Ind Eng Chem Prod Res Dev 9:55354. Kawahara FK, Johnson (1968) Proceedings of the 11th Conference Great Lakes

Research p 55055. Mattson JS, Mark HB (1970) Anal Chem 42:23456. Mattson JS, Mark HB (1969) Environ Sci Technol 3:16157. Kawahara FK (1972) J Chromatogr Sci 10:62958. Mark HB Jr, Yu T-C, Mattson JS, Kolpac R (1972) Environ Sci Technol 6:83359. Baier RE (1972) J Geophys Res 77:506260. Pierre LJ (1973) Applied Optics 12:203561. Lynch PF, Brown CW (1973) Environ Sci Technol 1:112362. Ahmadjian M, Baer CD, Lynch PF, Brown CW (1971) Environ Sci Technol 10:77763. Mattson JS, Mattson CS, Spencer MJ, Starks SA (1977) Anal Chem 49:29764. Kawahara FE (1969) Laboratory guide for the identification of petroleum products.

Department of the Interior Federal Water Pollution Central Administration, AnalyticalQuality Central Laboratory, Cincinnati, OH, USA

65. Institute of Petroleum Standardization Committee (1970) Journal of the Institute ofPetroleum 56:107

66. Ramsdale SJ, Wilkinson RE (1968) Journal of the Institute of Petroleum 54:32667. Brunnock JV (1968) Journal of the Institute of Petroleum 54:31068. Duckworth DF (1971) Aspects of petroleum pollutant analysis. In: Hepple P (ed)

Water pollution in oil. Institute of Petroleum, London, p 16569. Blumer M (1970) Mar Biol 5:19670. Adams IM (1967) Process Biochem 2:3371. Erhardt M, Blumer M (1972) Environ Pollut 3:17972. Zafiron OC, Myers J, Freestone F (1973) Mar Pollut Bull 4:8773. McKay TR (1972) In: Proceedings of the 9th International Symposium on GM

Chromatography, Montreux, Switzerland, October, pp 33–3874. Adland ER, Creaser LF, Matthews PHD (1972) Anal Chem 46:6475. Adlard ER, Matthews PHD (1971) Natural Physical Science 233:8376. Boylan DB, Tripp BN (1971) Nature 230:4477. Bulton NJ, Morris BF, Sas J, Bermuda J (1973) Biological Research, Special Publication

No. 1078. Zafiron OC, Oliver C (1973) Anal Chem 45:95279. Hertz HS, May WE, Chester SN, Gump BH (1976) Environ Sci Technol 10:900

References 429

80. Garra ME Jr, Muth J (1974) Environ Sci Technol 8:24981. Wasik SP, Brown RL (1973) Determination of hydrocarbon solubility in seawater and

the analysis of hydrocarbons in water-extracts. In: Proceedings of the conference onprevention and control of oil spills. American Petroleum Institute, Washington, DC,pp 223–237

82. Rasmussen DV (1976) Anal Chem 48:156283. Smith DH (1972) Anal Chem 44:53684. Walker JD, Colwell RP, Hamming MC, Ford HI (1975) Bull Environ Contam Tox 13:24585. Brown RA, Huffman (1976) Science 101:84786. Albaiges J, Albrecht P (1979) Int J Environ Anal Chem 6:17187. Erhardt M, Blumer M (1972) Environmental Pollution 3:17988. Done JM, Reid WK (1970) Separ Sci 5:82589. Witmer FE, Goilan A (1973) Environ Sci Technol 7:94590. Matsumoto G, Hanya T (1981) Water Res 15:21791. Schuldiner JA (1951) Anal Chem 23:167692. Lieberman M (1973) Report no. EPA-RZ-73-102. US Environmental Protection Agency93. Wilson CA, Ferreto EP, Coleman HJ (1975) American Chemical Society, Division of

Chemistry Reports 20:61394. Harwood JE, Huyser DJ (1968) Water Res 2:63195. Treguer P, Le Corre P, Courtot P (1972) J Mar Biol Assoc of the UK 52:104596. Antia NJ, McAllister CD, Parsons TR, Stephens K, Strickland JDH (1963) Limnol

Oceanogr 8:16697. Shah NM, Fogg GE (1973) J Mar Biol Assoc of the UK 53:32198. Shah NM, Wright RT (1974) Mar Biol 24:12199. Stradomskya AG, Goncharova IA (1967) Gidrokhim Mater 43:57

100. Garrett WD, Timmons CO, Jarvis NL, Kagarise RE (1963) Constitution and surfacechemical properties of sea slicks, part 1, Bay of Panama. Report no. 5925. US NavalResearch Laboratory, Washington, DC

101. Slowey JF, Jeffrey LM, Hood DW (1962) Geochim Cosmochim Acta 26:607102. Williams PM (1965) Journal of the Fisheries Research Board of Canada 22:1107103. Jeffrey LM (1966) J Am Oil Chem Soc 43:211104. Ackman RG, Sipos JC, Tocher CS (1967) Journal of the Fisheries Research Board of

Canada 24:635105. Mahadevan V, Stenroos L (1967) Anal Chem 39:1652106. Blumer M (1970) Dissolved organic compounds in seawater. Saturated and olefinic

hydrocarbons and singly branched fatty acids. In: Hood DW (ed) Organic matter innatural waters. Publication no. 1. Institute of Marine Science, University of Alaska,pp 153–167

107. Stauffer TB, MacIntyre WG (1970) Chesapeake Science 11:216108. Quinn JG, Meyers PA (1971) Limnol Oceanogr 16:129109. Meyers PA (1976) Limnol Oceanogr 21:315110. Garrett WD (1967) Deep-Sea Res 14:221111. Barger WR, Garrett WD (1970) J Geophys Res 75:4561112. Duce RA, Quinn JG, Olney CE, Piotrowicz SR, Ray BJ, Wade TL (1972) Science 176:161113. Schultz DM, Quinn JG (1972) Journal of Fisheries Research 29:1482114. Marty JC, Saliot A (1974) Journal Rech Atmosphere 13:563115. Ackman RG (1972) J Chromatogr Sci 10:560116. Di Corcia A, Samperi F (1974) Anal Chem 46:140117. Bjork RG (1975) Anal Biochem 63:80


118. White WR, Leenheer JA (1975) J Chromatogr Sci 13:386119. Sever J, Parker PL (1969) Science 164:1052120. Schultz DM, Quinn JG (1974) Mar Biol 27:143121. White RH, Miller SL (1976) Science 193:885122. Mueller HF, Larson TE, Ferretti M (1960) Anal Chem 32:687123. Koyama T, Thompson TG (1964) Journal of the Oceanographic Society of Japan

20:209124. Maksimova IV, Pimenova MN (1969) Mikrobiology 37:77125. Lamar WL, Goerlitz DF (1966) Organic acids in naturally colored surface waters.

Water supply paper no. 1817-A:1-17. Geological Survey126. Horikawa K (1972) Buseki Kagaku 21:806127. Goncharova IA, Khomenko AN (1971) Gidrokhim Mater 53:36 Ref: Zhur Khim 199D

(1970) Abstract No. 59339128. Borch RF (1975) Anal Chem 47:2437129. Hoffman NW, Ligo JC (1976) Anal Chem 48:1104130. Kieber DJ, Vaughan GM, Mopper K (1988) Anal Chem 60:1654131. Xiao HY, Lee C, Scranton MI (1993) Anal Chem 65:857132. Kamata E (1966) Bull Chem Soc Japan 39:1227133. Eberhardt MA, Sieburth JM (1985) Mar Chem 17:199134. Afghan BK, Kulkarni AU, Leung R, Ryan JF (1974) Environ Lett 7:53135. Zika RG (1977) An Investigation in marine photochemistry. PhD Dissertation.

Dalhousie University136. Belman S (1963) Anal Chim Acta 29:120137. Slawinska D, Slawinski J (1975) Anal Chem 47:2101138. Afghan BK, Kulkarni AU, Ryan JF (1975) Anal Chem 47:488139. MacKinnon MD (1977) The analysis of the total organic carbon in seawater: a. Devel-

opment of methods for the quantification of TOC b. Measurement and examinationof the volatile fraction of the TOC. PhD Dissertation. Dalhousie University

140. Corwen JP (1970) Volatile organic materials in seawater. In: Hood W (ed) Organicmatter in natural waters. Publication no. 1. Institute of Mar Sci, University of Alaska,pp 169–170

141. Alekseeva LM (1972) Trudy gos Oceanogr Inst 60:17142. Goulden PD, Brooksbank P, Day MB (1973) Anal Chem 45:2430143. Schlungbaum G, Behling A (1974) Acta Hydroch Hydrob 2:423144. Stilinovic L, Mujko II, Vukic B (1974) Arh Hig Rada Tokskko 25:247145. Gales ME Jr (1975) Analyst 100:841146. Wexler AS (1963) Anal Chem 35:1936147. Fountaine JE, Joshipura PB, Keliher PN, Johnson JD (1975) Anal Chem 46:62148. Boyd TJ (1994) J Chromatogr 662:281149. Katase T (1972) Kogyo Josui 1972:19150. Corcoran EG (1978) Environ Heath Persp 3:13151. Giam CS, Chan HS, Neff GS (1975) Anal Chem 13:2225152. Hites RA (1973) Environ Heath Persp 1973:17153. Lewis GJ, Rakestraw NW (1955) J Mar Res 14:253154. Josefsson B, Koroleff F (1976) Polyphenolic substances. In: Grasshoff K (ed) Methods

of seawater analysis. Verlag Chemie, New York, p 220–227155. Collier A (1958) Limnol Oceanogr 3:33156. Josefsson B, Uppstrom L, Ostling G (1972) Deep-Sea Res 19:385157. Mopper K, Gindler EM (1973) Anal Chem 56:440

References 431

158. Johnson KM, Sieburth JMcN (1977) Mar Chem 5:1159. Burney CM, Sieburth JMcN (1977) Mar Chem 5:15160. Hirayama H (1974) Anal Chim 70:141161. Williams PM, Craigie JS (1970) Microdetermination of uronic acids and related

compounds in the marine environment. In: Hood DW (ed) Organic matter in naturalwaters. Institute of Marine Science, Publication no. 1. University of Alaska, pp 509–519

162. Mopper K (1977) Mar Chem 5:585163. Hicks SE, Carey FG (1968) Limnol Oceanogr 13:361164. Andrews P, Williams PJ, Leb (1971) J Mar Biol Assoc 51:11165. Whittaker JR, Vallentyne JR (1957) Limnol Oceanogr 2:98166. Degens ET, Reuter JH, Shaw KNF (1964) Geochim Cosmochim Acta 28:45167. Starikova ND, Yablokova OG (1972) Okeanologiya 12:431168. Starikova ND, Yablokova OG (1972) Okeanologiya 12:898169. Geller A (1975) Archives Hydrobiology 47:295170. Josefsson BD (1970) Anal Chim Acta 52:65171. Larsson K, Degens EG (1967) Microchim Acta 2:328172. Mopper K, Degens EG (1972) Anal Biochem 45:147173. Modzeleski JE, Laurie WA, Nagy B (1971) Geochim Cosmochim Acta 34:825174. Tesarik K (1972) J Chromatogr 65:295175. Eklund G, Josefsson B, Roos C (1977) J Chromatogr 142:575176. Williams PJ, Le B (1964) Limnol Oceanogr 9:138177. Sinel’nikov VE, Breusov NG (1973) Trudy Ins. Biol V nutr Vod, Akad Nauk SSSR 212178. Afghan BK, Belliveau PE, Larose RH, Ryan JF (1974) Anal Chim Acta 71:355179. Guilbault GG, Kramer DN, Hackley E (1966) Anal Chem 38:1897180. Goncharova IA, Khomenko AN, Shuk IP (1974) Gidrokhim Mater 61:100181. Chriswell CD, Kissinger LD, Fritz JS (1976) Anal Chem 47:1325182. Kunte H, Slemrova J (1975) Zeitung Wasserabwasser Forschung 8:176183. Kleverlaan NTM (1975) Chemishe Weekblad 71:13184. Cassidy RM, LeGay DS, Frei RW (1974) J Chromatogr Sci 12:85185. Wang LK, Pek SL (1975) Ing Engng Chem 14:308186. Le Bihan A, Courtot-Coupez J (1970) Bull Chem Soc France 1:406187. Hon-Nami H, Hanya T (1980) Water Res 14:1251188. Hon-Nami H, Hanya T (1978) J Chromatogr 161:205189. Wang LK, Yang JY, Wang MH (1973) Engineering Bulletin Purdue University

Engineering External Series 142:76190. Wang LK, Yang JY, Wang MH (1975) J Am Water Works Assoc 67:6191. Kozarac Z, Cosovic B, Branica M (1975) Mar Sci Commun 1:147192. Hagihara K (1972) Kogai Bunseki Shihkin 5:68193. Den Tankelaar WAH, Bergshoeff G (1969) Water Res 3:31194. Petts KW, Parkes D (1978) Determination of anionic detergents within a concentration

range 0–1 mg/l in sewage effluents and waters by autoanalyser. Water Research Centrereport TR75

195. Longwell J, Maniece WD (1955) Analyst 80:167196. Abbott DC (1962) Analyst 87:286197. Södergren A (1966) Analyst 91:113198. Favretto L, Tunis F (1976) Analyst 101:198199. Bhat SR, Crisp PT, Eckert JM, Gibson MA (1981) Anal Chim Acta 116:191200. Crisp PT, Eckert JM, Gibson NA (1975) Anal Chim Acta 78:391201. Courtot-Coupez J, Le Bihan A (1969) Anal Lett 2:211


202. Le Bihan A, Courtot-Coupez J (1977) Anal Lett 10:759203. Gagnon MJ (1979) Water Res 13:53204. Crisp PT, Eckert JM, Gibson NA (1975) Anal Chim Acta 87:97205. Takano S, Takasake S (1976) Yukagku 25:31206. Favretto L, Stancher B, Tunis F (1978) Analyst 103:955207. Favretto L, Stancher B, Tunis F (1979) Analyst 104:241208. Favretto L, Stancher B, Tunis F (1980) Analyst 105:833209. Courtot-Coupez J, Le Bihan A (1969) Anal Lett 2:567210. Courtot-Coupez J, Le Bihan A (1971) Riv Ital Sostanza Grasse 20:672211. Grieff O (1966) Anal Abstr 13:3650212. Crisp PT, Eckert JM, Gibson NA (1979) Anal Chim Acta 104:93213. Ferguson PL, Iden CR, Brownawell BJ (2000) Anal Chem 72:4322214. Eklund G, Josefsson B, Roos C (1978) J High Res Chromatog 1:34215. Eklund G, Josefsson B, Roos C (1977) J Chromatogr 142:575216. Grob K, Grob G (1970) J Chromatogr 125:471217. Grob K (1977) Chromatographie 10:181218. Kaiser KLE, Oliver BG (1976) Anal Chem 48:2207219. Bellar TA, Lichtenberg JJ, Eichelberger JW (1976) Environ Sci Technol 10:926220. Grob K (1973) J Chromatogr 84:255221. Grob K, Zurcher F (1976) J Chromatogr 117:285222. Grob K, Grob G (1974) J Chromatogr 90:303223. Grob K, Grob G (1975) J Chromatogr 106:249224. Westerdorf RG (1982) International Laboratory September:32225. Waggot A, Reid WJ (1977) In: Separation of non-volatile organic compounds in

sewage effluent using high-performance liquid chromatography. Part 1. Technicalreport no. 52. Water Research Centre, Medmenham, UK

226. Bullister JL, Weiss RF (1988) Deep-Sea Res 35:839227. Dawson R, Riley JP, Tennant RH (1976) Mar Chem 4:83228. Lovelock JE, Maggs RJ, Wade RL (1973) Nature 241:194229. Lovelock JE (1975) Nature 256:193230. Murray AJ, Riley JP (1973) Nature 242:37231. Murray AJ, Riley JP (1973) Anal Chim Acta 65:261232. Kristiansen NK, Froeshuag M, Aune KT, Becher G, Lundanes E (1994) Environ Sci

Technol 28:1669233. Friant SL, Suffet JH (1979) Anal Chem 51:2161234. Cowen WF, Cooper WJ, Highfill JW (1975) Anal Chem 47:2483235. Gjavotchanoff S, Luessem H, Schlimme K (1971) Gas-u-Wasser Fach 112:448236. Heydrickx A, Peteghem C (1975) Eur J Toxicol 8:275237. Chesler SN, Gump BH, Hertz HP (1978) Anal Chem 50:805238. Dowty B, Green L, Laseter JL (1976) J Chromatogr Sci 14:187239. Waggot A (1978) In: Proceedings of the Analytical Division of the Chemical Society

15:232240. Malter L, Vreden V (1978) Forum Städte-Hygiene 29:37241. Luessen H (1972) Detection of trace materials by concentration by pressure

distillation. Tech Essen Vortrags Verhöff 62242. Jeüttner F (1994) Chemosphere 29:191243. Kasthurikrishnan N, Cooke RG (1995) Talanta 42:1325244. Hashimoto S, Matsuda M, Wakimotot T, Tatsukawa R (1995) Chemosphere 30:1979245. Park K, Williams WT, Prescott JM, Hood DW (1962) Science 138:331

References 433

246. Riley JP, Segar DA (1970) J Mar Biol Assoc 50:713247. Hosaku K, Maita Y (1971) Journal of the Oceanographic Society of Japan 27:27248. Starikova ND, Korzhikova RI (1969) Oceanogiya 9:509249. Lee C, Bada JL (1975) Earth Planet Sci Lett 26:61250. Zlobin VS, Perlyuk MF, Orlova TA (1975) Okeanologiya 15:643251. Semenov AD, Ivleva IN, Datsko VG (1964) Trudy Komiss, Anal Khim Akad Nauk SSSR

13:62252. Coughenower DD, Curl HC Jr (1975) Limnol Oceanogr 20:128253. Pocklington R (1971) Nature 230:374254. Bajor M, Bohling HZ (1970) Oceanographica Acta 249:190255. Froelich PM, Murphy LD (1977) Anal Chem 49:1606256. Whiteledge TE, Dugdale RC (1972) Limnol Oceanogr 17:309257. Hammer KD, Luck S (1987) Fresen Z Anal Chem 327:518258. Litchfield CD, Prescott JM (1970) Limnol Oceanogr 15:250259. Roth M (1971) Anal Chem 43:880260. Bayer E, Grom E, Kaltenegger B, Uhmann R (1976) Anal Chem 48:1106261. Roth M, Hampai A (1973) J Chromatogr 83:353262. Felix AM, Terkelsen G (1973) Arch Biochem Biophys 157:177263. Gardner WS (1975) Mar Chem 6:15264. Dawson R, Pritchard RG (1978) Mar Chem 6:27265. Mopper K (1986) Sci Total Environ 49:115266. Clarke ME, Jackson GA, North WJ (1972) Limnol Oceanogr 17:749267. Palmark KH (1963) Acta Chim Scand 17:1456268. Siegal A, Degens ET (1966) Science 151:1098269. Webb KL, Wood L (1967) Improved techniques for analysis of free amino acids in

seawater. In: Scova NB et al. (eds) Automation in analytical chemistry. TechniconSymposium 1966, Vol. 1. Mediad Inc., New York pp 440–444

270. Hobbie JE, Crawford CC, Webb KL (1968) Science 159:1463271. Brockman UH, Eberlein K, Junge HD, et al. (1974) Entwicklung natürlicher

Plankton-Populationen in einem outdoor-tank mit nahrstoffarmen Meerwasser. II.Konzentrationsveränderungen von gelösten neutralen Kohlen Meeresforschung. Vol.6. Springer-Verlag, Berlin

272. Williams PJ, Le B, Berman T, Holm-Hansen O (1976) Mar Biol 35:41273. Bohling H (1970) Mar Biol 6:213274. Bohling H (1972) Mar Biol 16:281275. Garrasi C, Degens ET (1976) Analytische Methoden zur säulenchromatographischen

Bestimmung von Aminosauren und Zuckern im Meereswasser und Sediment.Berichte aus dem Projekt DFG-De74/3: “Litoralforschung-Abwasser in Küstennähe”.DFG-Abschlusskolloquium, Bremerhaven

276. Tatsumoto M, Williams WT, Prescott JM, Hood DW (1961) J Mar Res 19:89277. Pocklington R (1970) A new method for the determination of amino-acids in sea

water and an investigation of the dissolved free amino-acids of North Atlantic Oceanwaters. PhD Dissertation. Dalhousie University

278. Pocklington P (1972) Anal Biochem 45:403279. Gehrke CW, Lamkin WM, Stalling DL, Shahrokhi F (1965) Biochem Biophys 19:328280. Gardner WS, Lee GE (1973) Environ Sci Technol 7:719281. Debeika EV, Fyabov AK, Nabivanets BI (1973) Gidrobiol Zhur 9:109282. Udenfriend S, Stein S, Bohlen P, Diarman W, Leimgruber W, Weigele M (1972) Science

178:871


283. North BB (1975) Limnol Oceanogr 20:20284. Parkard TT, Dortch Q (1975) Mar Biol 33:347285. Baier RM, Goupil DW, Perimutter S, King R (1974) Journal Rech Atmosphere 8:1286. Pillai TNV, Ganguly AK (1970) Current Science 39:501287. Hicks E, Riley JP (1980) Anal Chim Acta 116:137288. Strickland JDH, Solorzano L (1966) Determination of monoesterase hydrolysable

phosphate and phosphomonesterase activity in seawater. Some continued studies inmarine science. Springer, Berlin, pp 665–674

289. Maeda M, Taga N (1973) Mar Biol 20:58290. Batley GE, Florence TM, Kennedy JR (1973) Talanta 20:987291. Di Corcia A, Samperi F (1974) Anal Chem 46:977292. Varney MS, Preston MR (1985) J Chromatogr 348:265293. Petty RL, Michel C, Snow JP, Johnson KS (1982) Anal Chim Acta 142:299294. Yang XH, Lee C, Scranton MI (1991) Anal Chem 65:572295. Florence TM, Farner YJ (1973) Anal Chim Acta 63:255296. Hashimoto A, Sakino H, Yamagami E, Tateishi S (1980) Analyst 105:787297. Hashimoto A, Kozima T, Sakino H, Ankey T (1979) Water Res 13:509298. Barshick S-A, Griest WH (1998) Anal Chem 70:3015299. Shinohara R, Kido A, Okamoto Y, Takeshita R (1983) J Chromatogr 256:81300. Junk GA, Richard JJ, Grieser D, Witiak D, Witiak JL, Arguello MD, Vick R, Svec HJ,

Fritz JS, Calder GV (1974) J Chromatogr 99:745301. Emmet RT (1969) Spectrophotometric determination of urea in natural waters with

hypochlorite and phenol. NAVSHIPRANDLAB Annapolis Report 2663302. Newell BS, Morgan B, Cundy J (1967) J Mar Res 25:201303. Muravskaya ZA (1973) Gidrobiol Zhur 9:127304. DeMarche JM, Curl H Jr, Coughenower DD (1973) Limnol Oceanogr 18:686305. Mitamura O, Osaka K (1987) Daigaku Kiyo Dai-3-bumon 36:159306. Von Breymann, de Angelis MA, Gordon LI (1982) Anal Chem 54:1209307. Brown L, Rhead M (1979) Analyst 104:391308. Stolzberg PJ (1977) Anal Chim Acta 92:139309. Lovelock JE, Maggs RJ, Rasmussen RA (1972) Nature 237:452310. Yamavka V, Tanimoto T (1976) J Agr Chem Soc 50:147311. Leck C, Baagander LE (1988) Anal Chem 60:1680312. Charlson RJ, Lovelock JE, Andreae MO, Warren SG (1987) Nature 326:655313. Malin G, Turner S, Liss PS, Holligan P, Harbour D (1993) Deep-Sea Res 40:1487314. Turner SM, Malin G, Liss PS, Harbour DS, Holligan PM (1988) Limnol Oceanogr 33:364315. Turner SM, Malin G, Liss PS (1989) In: Saltzmann E, Cooper WJ (eds) Biogenic sulfur

in the environment. ACS symposium series no. 393. American Chemical Society,Washington, DC, pp 183–200

316. Holligan PM, Turner SM, Liss PS (1987) Cont Shelf Res 7:213317. Keller MD, Bellows WK, Guillard RRL (1989) In: Saltzmann E, Cooper WJ (eds)

Biogenic sulfur in the environment. ACS symposium series no. 393. AmericanChemical Society, Washington, DC, pp 167–182

318. Stefels J, van Boekel WHM (1993) Mar Ecol Prog Ser 97:11319. Challenger F, Simpson MI (1948) J Chem Soc 3:1591320. Ledyard KM, Dacey JWH (1994) Mar Ecol Prog Ser 110:95321. Smith GC, Clark T, Knutsen L, Barrett E (1999) Anal Chem 71:5563322. Shea D, McCrehan WA (1988) Anal Chem 60:1449323. Duare LM, Stock JT (1978) Anal Chem 50:1891

References 435

324. Andreae MO (1980) Anal Chem 52:150325. Andreae MO (1977) Anal Chem 49:820326. CapitanF,AlonsoE,AvidadR,Capitan-VallveyLF,Vilchez JL (1993)AnalChem65:1336327. Van den Berg CMG, Househam BC, Riley JP (1988) J Electroanal Chem–Interfacial

Electrochemistry 239:137328. Turner SM, Nightingale PD, Broadgate W, Liss PS (1995) Deep-Sea Res Part II: Topical

Studies in Oceanography 42:1059329. Wanatabe S, Demura K, Tsunogai S (1987) Hokkaido Daigaku Suisan Gabuku Kenkyo

Iho 38:288330. Przyjazny A (1988) J Chromatogr A 346:61331. Savchuk SA, Rudenko BA, Brodskii ES, Soifer VS (1995) J Anal Chem 50:1081

(Translated from Zhur Anal Khim)332. Wilson AJ, Forester J (1974) US Environmental Protection Agency Report No.

EPA-600-7-74, p 108333. Wilson AJ, Forester J, Knight J (1969) US Fish Wildlife Service Circular 355, 18–20,

Centre for Estuaries and Research, Gulf Breeze, Florida, USA334. US Environmental Protection Agency Analytical Quality Control Laboratory, US

Government Printing Office, Washington DC (1972)335. Wilson AJ, Forester JJ (1971) Association of the Officers in Analytical Chemistry 54:525336. Picer IV, Picer M, Strohal P (1975) Bull Environ Contam Tox 14:565337. Wilson AJ (1976) Bull Environ Contam Tox 15:515338. Aspila KI, Carron JH, Chau ASY (1977) Journal of the Association of Official Analytical

Chemists 60:1097339. Analytical Methods Manual (1974) Inland Water Directorate Water Quality Branch

Ottawa Ontario Canada340. Miles JR, Tu CM, Harris OR (1969) J Econ Ent 62:1334341. Chau ASY, Rosen JD, Cochrane WB (1971) Bull Environ Contam Tox 6:225342. Chau ASY (1974) Journal of the American Oil and Colours Association 57:585343. Thompson JE (1974) Analysis of Pesticide Residues in Human and Environmen-

tal Samples, US Environmental Protection Agency, Research Triangle Park, USEnvironmental NC Sec 10A DG

344. Eichellberger JW, Lichtenberg JL (1971) Environ Sci Technol 5:541345. Samuelsen OB (1987) Aquaculture 60:161346. Zitko V, Choi P (1971) PCB and Other Industrial Halogenated Hydrocarbons in

the Environment. Fish Research Board Ottawa Canada Technical Report No. 272Biological Station St Andrews NB

347. Oller WL, Cramwer MF (1975) J Chromatogr Sci 13:296348. Edwards R (1970) Chem Ind – London 1340349. Fishbein L (1972) J Chromatogr 68:345350. Schulte E, Their HP, Acker L (1976) Deut Lebensmittel Rundschau 72:229351. Goke G (1975) Deut Lebensmittel Rundschau 71:309352. Ashling B, Jensen S (1970) Anal Chem 42:1483353. Musty PR, Nickless G (1974) J Chromatogr 89:185354. Girenko DB, Klisenko MA, Dishcholka YK (1975) Hydrobiological Journal 11:60355. Elder D (1976) Mar Pollut Bull 7:63356. Harvey GR (1973) Report of the US Environment Protection Agency, EDA-R2-73-177357. Picer N, Picer M (1980) J Chromatogr 193:357358. Holden AV, Marsden K (1969) J Chromatogr 44:481359. Snyder DE, Reinert RE (1971) Bull Environ Contam Tox 6:385


360. Picer M, Abel M (1978) J Chromatogr 150:1191361. Hernandez Garcia J, Sosa Ferrera Z, Bermejo Martin-Lázaro AJ, Santana Rodriguez

JJ (1994) Anal Lett 27:1355362. Hernandez Garcia J, Betancort Rodriguez JR, Bermejo Martin-Lázaro AJ, Santana

Rodriguez JJ (1994) Anal Chim Acta 290:146363. Petrick G, Schulz DE, Duinker JC (1988) J Chromatogr 435:241364. Pedersen-Bjergaard S, Semb SI, Vedde J, Brevik EM (1996) Chromatographia 43:44365. Leoni V, Pucelti G, Grella AJ (1975) J Chromatogr 106:119366. Leoni V (1971) J Chromatogr 62:63367. Johnston LV (1965) Journal of Association of Official Analytical Chemists 48:668368. Claeys RR, Inman RD (1974) Journal of Association of Official Analytical Chemists

57:399369. Leoni V, Pucetti G (1969) J Chromatogr 43:388370. Leoni V, Pucetti G, Columbo RJ, Oviddo O (1976) J Chromatogr 125:399371. Millar JD, Thomas RE, Schatterberg HJ (1981) Anal Chem 53:214372. Kelly AG, Cruz I, Wells DE (1993) Anal Chim Acta 276:3373. Boyd TJ (1994) J Chromatogr 662:281–92374. Sullivan KF, Atlas EL, Giam C-S (1981) Anal Chem 53:1718375. Petrick G, Sculz-Bull DE, Martens V, Scolz K, Duinker JC (1996) Mar Chem 54:97376. Minear RA (1972) Environ Sci Technol 6:431377. Minear RA, Walanski KA (1974) Investigation of the chemical identity of sol-

uble organophosphorus compounds found in natural waters. Research reportUILU-WRC-74-0086

378. McElvroy WD, Strehler BL (1949) Arch Biochem 22:420379. Wildish DJ (1976) Determination of adenosine 5′-triphosphate in estuarine water and

sediments by firefly bioluminescence assay. Technical report of the fisheries maritimeservice research division, no. 649

380. Hofer-Siegrist L, Schweiz A (1976) Hydrology 38:49381. Weber K (1976) Water Res 10:237382. Lores EM, Moore JC, Knight J (1985) J Chromatogr Sci 23:124383. Daughton CG, Crosby DC, Garnes RL, Hsieh DPH (1976) J Agr Food Chem 24:236384. Michel HO, Gordon EC, Epstein J (1973) Environ Sci Technol 7:1045385. Cembella AD, Antia NJ (1986) Mar Chem 19:205386. Ahern F, Eckert JM, Hain SF, Leggett KEA, Payne NC, Williams KL (1987) Anal Chim

Acta 199:259387. Grift N, Lockhart WL (1974) Journal of the Association of Official Analytical Chemists

57:1282388. Wu TL, Lambert L, Hastings D, Banning D (1980) Bull Environ Contam Tox 24:411389. Abel M, Evans KM, Fileman TW, Mantoura RFC (1992) Anal Chim Acta 268:195390. Durand G, Barcelo D (1993) Talanta 40:1665391. Molina C, Durand G, Barcelo D (1995) J Chromatogr A 712:113392. Barcelo D, Durand G, Bouvet V, Nielen M (1973) Environ Sci Technol 27:271393. Durand G, Chiron S, Bouvet V, Barcelo D (1992) Int J Environ Anal Chem 49:31394. Ferrer I, Barcelo D (1999) J Chromatogr A 854:197395. Paradis M, Ackman RG (1977) Journal of the Fisheries Research Board, Canada 34:2156396. Lewis RW (1969) Limnol Oceanogr 14:35397. Lee RF, Nevenzei JC, Pattenhofer GA (1971) Mar Biol 9:99398. Volimirov B (1982) Mar Ecol 3:97399. Olsen CR, Cutshall NH, Larsen TL (1982) Mar Chem 11:501

References 437

400. Whittle KJ, Hardy R, Mackie PR, McGill AS (1982) Philos T Roy Soc B 297:193401. Simkiss K (1983) Journal of the Marine Biological Association, UK 63:1402. Florence TM, Lumsden BG, Fardy J (1983) Anal Chim Acta 151:281403. Parrish CC, Ackman RG (1983) (1983) J Chromatogr A 262:103404. Delmas RP, Parrish CC, Ackman RG (1984) Anal Chem 56:1272405. Wun CK, Walker RW, Litsky W (1976) Water Res 10:955406. Dutka BT, Chau ASY, Coburn J (1974) Water Res 8:1047407. Wun CK, Walker RW, Litsky W (1978) Health Lab Sci 15:67408. Adams DD, Richards RA (1960) Deep-Sea Res 15:471409. Belyaeva AN (1974) Okeanologiya 14:1101410. Matthews SW, Smith LL (1968) Lipids 3:239411. Smith LL, Gouron RE (1969) Water Res 3:141412. Saliot A, Barbier M (1973) Deep-Sea Res 20:1077413. Saliot A, Barbier M (1973) Advances in Organic Geochemical Procedures Institute

Meeting. In: Proc Mtg US Geochemical Institute, pp 607–617414. Matsushima H, Wakimoto T, Tatsukawa R (1975) Japan Analyst 24:342415. Gagosian RB (1975) Geochim Cosmochim Acta 39:1443416. Attaway D, Parker PL (1970) Science 169:674417. Kanazawa A, Teshima SI (1971) Journal of the Oceanographic Society of Japan 27:207418. Hanck KW, Dillard JW (1977) Anal Chem 49:404419. Stolzberg RJ, Rosin D (1977) Anal Chem 49:226420. Chau YK, Gachter R, Lum-Shue-Chan K (1974) Journal of the Fisheries Research

Board of Canada 31:1515421. Buffle J, Greter F-L, Haerdi W (1977) Anal Chem 49:216422. Jones DR, Mahahan SE (1975) Anal Lett 8:421423. Jones DR, Mahahan SE (1976) Anal Chem 48:502424. Kaiser KLE (1973) Water Res 7:1465425. Egan WG (1974) Journal of the Marine Technical Society 8:40426. Ghassemi M, Christman RF (1968) Limnol Oceanogr 13:583427. Gjessing ET (1970) Environ Sci Technol 4:437428. Brown W (1975) Mar Chem 3:253429. Wangersky PJ (1978) Production of dissolved organic matter. In: Marine Ecology (ed.

Kinne O), Vol. 4, John Wiley, New York430. Sieburth JMcN, Jensen A (1968) J Exp Mar Biol Ecol 2:174431. Sieburth JMcN, Jensen A (1969) J Exp Mar Biol Ecol 3:275432. Pocklington R, Hardstaff WR (1974) Journal of the International Conference on

Exploration of the Sea 36:92433. Stoeber H, Eberle SH (1974) Organic Acids in the Rhine and Some of its Tributaries.

Kernforschungszentrum Karlsruhe, KFK 1969 UF, pp 58–81434. Revina SK, Kriul’kov VA (1972) Trudy Gos Oceanographic Institute 113:42435. Bilikova A (1973) Vod Hospod B 23:44436. Kloster MB (1975) J Am Water Works Assoc 66:44437. Bilikova A (1974) Vod Hospod B 24:20438. Eberle SH, Hoesle C, Krueckeberg C (1974) New method for simultaneously

determining lignosulphonic acid and humic acid in water, using differential pulsepolarography. Kernforschungszentrum Karlsruhe, KFK 1969 UF, pp 44–57

439. Paluch J, Stangret S (1972) Fortschr Wasserchem Ihrer Grenzgeg 14:89440. Pocklington R, MacGregor CD (1973) Int J Environ Anal Chem 3:81441. Almgren T, Josefsson B (1973) Sveridge Papperstidn 76:19


442. Wilander A, Kvarnas H, Lindel T (1974) Water Res 8:1037443. Almgren T, Josefsson B, Nyquist G (1975) Anal Chim Acta 78:411444. Kononova MM (1966) In: Soil Organic Matter Pergamon Press New York445. King LH (1967) Isolation and Characterization of Organic Matter from Glacial-Marine

Sediments on the Scotian Shelf. Report BIO 67-4, pp 1–18446. Rashid MA, King LH (1969) Geochim Cosmochim Acta 33:147447. Pierce RH Jr, Felbeck GT Jr (1972) A comparison of three methods of extracting

organic matter from soils and marine sediments. In: Povoledo D, Golterman HL(eds) Humic Substances. Center for Agriculture Publication and Documentation,Wageningen, Netherlands, pp 217–232

448. Stuermer DH, Harvey GR (1974) Nature 250:480449. Stuermer DH, Harvey GR (1977) Deep-Sea Res 24:303450. Mantoura RFC, Riley JP (1975) Anal Chim Acta 76:97451. Stuermer DH, Payne JR (1976) Geochim Cosmochim Acta 40:1109452. Schnitzer M (1972) Chemical, spectroscopic, and thermal methods for the classi-

fication and characterization of humic substances. In: Povoledo D, Golterman HL(eds) Humic Substances. Center for Agriculture Publication and Documentation,Wageningen, Netherlands, pp 293–310

453. Salfeld JC (1972) Ultraviolet and visible absorption spectra of humic systems.In Humic Substances (eds Povoledo D, Golterman HL), Center for AgriculturePublications and Documentation, Wageningen, The Netherlands, p. 269–280

454. Generalova VA (1974) Gidrokhim Mater 60:186455. Sil’chenko SA, Rubinshtein RN, Vasileva LN (1972) Zhur Analytical Khim 27:2465456. Manka J, Rebhun M, Mandelbaum A, Bortinger A (1974) Environ Sci Technol 8:1017457. Schnitzer M, Skinner SIM (1974) Can J Chem 52:1072458. Skinner SIM, Schnitzer M (1975) Anal Chim Acta 75:207459. Strickland JDH, Parsons TR (1972) A Practical Handbook of Seawater Analysis. 2nd

edition, Bulletin of the Fisheries Research Board of Canada, No. 067460. Strickland JDH, Parsons TR (1972) Determination of soluble organic carbon. In:

Goldberg ED (ed) A guide to marine pollution. Gordon and Breach Science Publishers,New York, pp 70–76

461. Rai H (1973) Theoretical Angerolimnology 18:1864462. Messiha-Hanna RG Okeanologiya 13:704463. Quirry MR, Holland WE, Waring RC, Hale GM (1975) High sensitivity laser absorption

spectroscopy of laboratory aqueous solutions and of natural Missouri waters. PBreport no. 243123, US National Technical Information Service, Springfield, VA, USA

464. Bjarnborg B (1979) Report No. 1239, P22ABJA, Swedish Environmental ProtectionBoard, Solna, Sweden

465. Boto KG, Bunt TS (1978) Anal Chem 50:392466. Murray AP, Gibbs CF, Longmore AR, Flett DJ (1986) Mar Chem 19:211467. Abayashi JK, Riley JP (1979) Anal Chim Acta 107:1468. Garside C, Riley JP (1969) Anal Chim Acta 46:179469. Karabashev GS (1992) Oceanographia Acta 15:255470. Ryther JH, Guillard RRL (1962) Can J Microbiol 8:437471. Ohwada K, Otshhata M, Taga N (1972) Bulletin of the Japanese Society of Fish Science

28:817472. Gold K (1964) Limnol Oceanogr 9:343473. Carlucci AF, Silbernagel SB (1966) Can J Microbiol 12:175474. Natarajan KV, Dugdale RC (1966) Limnol Oceanogr 11:621

References 439

475. Natarajan KV (1968) Appl Microbiol 16:366476. Litchfield CD, Hood DW (1965) Appl Microbiol 13:886477. Ohwada K (1972) Mar Biol 14:10478. Carlucci AF, Silbernagel SB (1967) Can J Microbiol 13:979479. Sharma GM, DuBois HR, Pastore AT, Bruno SF (1979) Anal Chem 51:196480. Suzucki T, Mitsuya T, Matsubara H, Yamasaki M (1998) J Chromatogr A 815:155481. Vastano SE, Milne PJ, Stahoec WL, Mopper K (1987) Anal Chim Acta 201:127482. Brown RA, Searl TD, Elliott JJ et al. (1973) Distribution of heavy hydrocarbons in

some Atlantic waters. In: Proceedings of the conference on prevention and control ofoil spills. American Petroleum Institute, Washington, DC, pp 509–519

483. Hites RA, Biemann K (1972) Science 178:158484. Hites RA, Biemann WG (1975) Identification of specific organic compounds in a highly

anoxic sediment by gas chromatographic-mass spectrometry and high resolutionmass spectrometry. In: Gibb RP Jr (ed) Analytical methods in oceanography.American Chemical Society, Washington, DC, pp 188–201

485. Mengerhauser JV (1973) Crystal microbalance for determination of oil in water. ADreport no. 762107US. National Technical Information Service, Springfield, VA, USA

486. Bieri RH, Walerk AL, Lewis BW et al. (1974) Identification of hydrocarbons in anextract from estuarine water accommodated no. 2 fuel oil. Special publication no.409. National Bureau of Standards, Washington, DC, p 149

487. Hertz HS, Chesler SN, May WE et al. (1974) Methods for trace organic analysis insediments and marine organisms. Special publication no. 409. National Bureau ofStandards, Washington, DC, p 197

488. Cretney WJ, Johnson WK, Wong CS (1977) Trace analysis of oil in sea water byfluorescence spectroscopy. Unpublished manuscript. Pacific Mar Sci Report, pp 77–75

489. Barbier M, Joly D, Saliot A, Tourres D (1973) Deep-Sea Res 20:305490. Zobova NA (1973) Okhr Ikh Resur 4:45491. Matsushima H, Hanya T (1975) Bunseki Kagaku 24:505492. Maxmanidi ND, Kovaleva GI, Zobova NS (1974) Okeanologiya 15:453493. Hellman H (1976) Zeitschrift fur Analytische Chemie 278:263494. Hughes DR, Belcher RS, O’Brien EJ (1973) Bull Environ Contam Tox 10:170495. Majori L, Petronio F, Nedoclan G (1973) Ig Mod 66:150496. Ahmed SM, Beasley MD, Efromson AC, Hites RA (1974) Anal Chem 46:1858497. Brown RA, Elliott JJ, Searl TD (1974) Measurement and characterisation of non-

volatile hydrocarbons in ocean water. Special publication no. 409. National Bureau ofStandards, Washington, DC, p 131

498. Mallevialle J (1974) Water Res 8:1071499. Brown RA, Elliott JJ, Kelliher JM, Searl TD (1975) Sampling and analysis of non-

volatile hydrocarbons in ocean water. In: Gibb RP Jr (ed), Analytical Methods inOceanography, American Chemical Society, Washington, pp 172–187

500. LeRoux JH (1975) Instrument for detecting and measuring trace quantities ofhydrocarbons in water. South African Patent No. 02.379 (C1 BOID GOIF)

501. Zsolnay A (1973) Chemosphere 2:253502. Zsolnay A (1974) Determination of aromatic and total hydrocarbon content in

submicrogram and microgram quantities in aqueous systems by means of high-performance liquid chromatography. Special publication no. 409. National Bureau ofStandards, Washington, DC, p 119

503. Zsolnay A (1974) J Chromatogr 90:79


504. Ellis DW (1979) Analysis of aromatic compounds in water using fluoresence andphosphorescence. PB report no. 212268. National Technical Information Service,Springfield, VA, USA

505. Mackie PR, Whittle KJ, Hardy R (1974) Mar Sci 2:359506. Simeoneit BR, Smith DH, Eglinton G, Burlingame AL (1973) Arch Environ Contam

Toxicol 1:193507. Blaylock JW, Bean RM, Wildung RE (1974) Determination of extractable organic

material and analysis of hydrocarbon types in lake and coastal sediments. Specialpublication no. 409. National Bureau of Standards, Washington, DC, p 217

508. Jeffrey LM, Rasby BF, Stevenson B, Hood DW (1964) Lipids of ocean water. In:Colombo U, Hobson GD (eds) Advances in organic geochemistry. Pergamon Press,Oxford, pp 175–197

509. Law RJ, Marchand M, Dahlmann G, Fileman TW (1987) Mar Pollut Bull 18:486510. Hiraide M, Ren FN, Tamura R, Mizuike A (1987) Mikrochimica Acta 416:137511. Husser ER, Stehl RH, Price DR, De Lap RA (1977) Anal Chem 49:154512. Tatsumo M, Williams TJ, Prescott JM, Hood DW (1961) J Mar Res 19:89513. Degens ET, Reuter JH, Shaw KNF (1964) Geochim Cosmochim Acta 28:47514. Rittenberg SC, Emery KO, Hulsemann J, Degens ET, Fay RC, Reuter LH, Grady JR,

Richardson SH, Bray EE (1963) Sediment Petrol 33:140515. Webb KL, Wood L (1967) Improved techniques for analysis of free amino acids in

seawater In: Automation in Analytical Chemistry (Technicon Symposium 1966) (edsScova NB et al.) New York Mediad Inc., Vol. 1, p 440–446

516. Crawford CC, Hobbie TE, Webb KL (1974) Ecology 55:551517. Brockmann UH, Eberlein K, Junge HD, Majer-Reimer E, Siebers D, Trageser H

(1974) Entwichlung natürlicher Planktonpopulationen in einem outdoor-Tankmit nährstoffarmem Meerwasser. II. Konzentrationsveränderungen von gelöstenneutralen Kohlenhydraten und freien gelösten Amino-säuren. Berichte aus demSonderforschungsbereich Meeresforschung. SFB94. Universitat Hamburg 6:166

518. Dawson R, Mopper K (1977) Anal Biochem 83:100519. Lee C, Bada JL (1977) Limnol. Oceanog 22:502520. Daumas RA (1976) Mar Chem 4:225521. Dawson R, Gocke K (1977) Oceanol Acta 1:80522. Chian ESK, Bruce WN, Fang HHP (1975) Environ Sci Tech Technology 9:52523. Klein E, Eichelberger J, Eyer C, Smith J (1975) Water Res 9:807524. Harvey GR (1972) Absorption of chlorinated hydrocarbon from seawater by

a crosslinked polymer, Woods Hole, MA, USA525. Sackmauereva M, Pal’usova O, Szokalay A (1977) Water Res 11:551526. Leoni V, Pucetti G, Grella A (1975) J Chromatogr 106:119527. Riley JP, Taylor D (1969) Anal Chim Acta 46:307528. Osterroht C (1972) Kiel Meerseforsch 28:48529. Sauer WA, Jadamec JR, Sagar RW, Killeen TJ (1979) Anal Chem 51:2180530. Chapman G, Rao AC (1967) Nature 214:627531. Bertsch W, Anderson E, Holzer G (1975) J Chromatogr 112:701532. Dravniecks A, Krotoszynski BK, Whitfield J, O’Donnell A, Burgwald T (1971) Environ

Sci Tech 5:1220533. Novak J, Zluticky J, Kubelka V, Mostecky J (1973) J Chromatogr 76:45534. The Oil Companies International Study Group for Conservation of Clean Air and

Water, Europe (1982) In: Analysis of trace substances in aqueous effluents frompetroleum refineries. Concawe report no. 6/82

References 441

535. Ochsenkuhn-Petsopolu M, Ochsenkuhn KM, Darusaikis G, Glannakis AN, Smith HD(1995) Can J Appl Spectros 40:66

536. Nagase T, Tao H, Imagawa T, Miyazaki A (1992) Shigen to Kankyo 1:263537. Schintu M, Kauri T, Contu A, Kudo A (1987) Ecotox Environ Safe 14:208538. Burgess RM (2000) Int J Environ Pollut 13:2539. Fuoco R, Ceccarini A, Onor M, Marrara L (1999) J Chromatogr 846:387540. Law RJ, Klingsoyr R (2000) Int J Environ Pollut 13:262541. Frank H, Cristoph EH, Holm-Hanson O, Bullister JL (2002) Environ Sci Technol 36:12542. Riu J, Eichhorn P, Guerrero JA, Knepper TP, Barcolo D (2000) J Chromatogr 889:221543. Metcalfe JS, Bell SG, Codd GA (2000) Water Res 34:2761

10 Organometallic Compounds

10.1Organoarsenic Compounds

Large amounts of arsenic enter the environment each year because of the use ofarsenic compounds in agriculture and industry as pesticides and wood preser-vatives. The main amount is used as inorganic arsenic (arsenite, arsenate), andabout 30% as organoarsenicals such as monomethylarsinate and dimethylarsi-nate. Arsenic is known to be relatively easily transformed between organicand inorganic forms in different oxidation states by biological and chemicalaction [1, 2]. Until recently, most of the analytical work has been concernedonly with the total content of arsenic. But as the toxicity and biological activityof the different species vary considerably, information about the chemical formis becoming of great importance in environmental analysis.

Arsenic being present in inorganic form and as a variety of organic com-pounds implies the use of analytical strategies which use physical and chemicalseparations, as well as general or specific detectors. Since the lower molecularweight organic arsenic compounds, as well as some inorganic arsenic com-pounds, are easily vapourised, both gas chromatography and direct vapourgeneration atomic absorption are favoured analytical methods.

A review of the analytic chemistry of arsenic in the sea, including occur-rence, analytical methods, and the establishment of analytical standards, hasbeen published [3].

Earlier investigations have shown that the major known organic arseniccompound in the environment is dimethylarsinate. For that reason, specificand sensitive methods for the determination of this compound are needed.Several methods exist for this determination, often together with other arsenicspecies. Hydride generation and selective vaporisation of cold-trapped arsinesin combination with various detection systems seem to be the methods mostfrequently used [4–6]. These methods are applicable to the species AsIII, AsV,monomethylarsinates, dimethylarsinate, and trimethylarsine oxide. The opti-mum conditions for generation of the respective arsine are, however, differentfor all these species with regard to the pH of the generating solution [7,8]. Thesemethods also suffer from interferences from numerous inorganic ions [9].Preconcentration in a toluene cold-trap following arsine generation was suc-

444 10 Organometallic Compounds

cessful for monomethyl arsinate and dimethyl arsinate, but nonquantitativerecoveries of dimethylarsinate were reported, probably because of the problemmentioned above with the arsine generation step. Molecular rearrangementsoccurring during arsine generations are, however, reported to be minimised ifsodium tetrahydroborate is introduced as a pellet [10]. To simplify the deter-mination of the different species, improved separation of the arsines using gaschromatography may be necessary.


The simplest analytical method is direct measurement of arsenic in volatilemethylated arsenicals by atomic absorption [11]. A slightly more complicatedsystem, but one that permits differentiation of the various forms of arsenic,uses reduction of the arsenic compounds to their respective arsines by treat-ment with sodium borohydride. The arsines are collected in a cold trap (liquidnitrogen), then vaporised separately by slow warming, and the arsenic is mea-sured by monitoring the intensity of an arsenic spectral line, as produced bya direct current electrical discharge [1, 12, 13]. Essentially the same methodwas proposed by Talmi and Bostick [10] except that they collected the arsinesin cold toluene (–5 ◦C), separated them on a gas chromatography column, andused a mass spectrometer as the detector. Their method had a sensitivity of0.25 µg/l for water samples.

Another variation on the method [4] with slightly higher sensitivity (severalng/l) used the liquid nitrogen cold trap and gas chromatography separation,but used the standard gas chromatography detectors or atomic absorption forthe final measurement. These workers found four arsenic species in naturalwaters.

Persson and Irgum [14] fulfilled a requirement to determine sub ppm lev-els of dimethyl arsinate by preconcentrating the organoarsenic compoundon a strong cation exchange resin (Dowex AG 50 W-XB). By optimising theelution parameters, dimethyl arsinate can be separated from other arseni-cals and sample components, such as group I and II metals, which can in-terfere in the final determination. Graphite furnace atomic absorption spec-trometry was used as a sensitive and specific detector for arsenic. The de-scribed techniqueallowsdimethylarsinate tobedetermined ina sample (20 ml)containing a 105-fold excess of inorganic arsenic with a detection limit of0.02 µg As/l.

The results show that good recoveries were obtained from artificial sea-waters, even at the 0–05 µg/l level, but for natural seawater samples therecoveries were lower (74–85%). This effect could be attributed to organicsample components that eluted from the column together with dimethyl arsi-nate.

10.1 Organoarsenic Compounds 445


In the method for [17] inorganic arsenic the sample is treated with sodiumborohydride added at a controlled rate (Fig. 10.1). The arsine evolved is ab-sorbed in a solution of iodine and the resultant arsenate ion is determined pho-tometrically by a molybdenum blue method. For seawater the range, standarddeviation, and detection limit are 1–4 µg/l, 1.4%, and 0.14 µg/l, respectively;for potable waters they are 0–800 µg/l, about 1% (at 2 µg/l level), and 0.5 µg/l,respectively. Silver and copper cause serious interference at concentrations ofa few tens of mg/l; however, these elements can be removed either by prelimi-nary extraction with a solution of dithizone in chloroform or by ion exchange.

The precision of the method was tested by carrying out replicate analyses(10) on 150 ml aliquots of two seawater samples from the Irish Sea. Mean (±sd)arsenic concentrations of 2.63 ± 0.05 and 2.49 ± 0.05 µg/l amounts of werefound. The recovery of arsenic was checked by analysing 150 ml aliquots ofarsenic-free seawater which had been spiked with known amounts of ar-senic (V). The results of these experiments shows that there is a linear re-lationship between absorbance and arsenic concentration and that arseniccould be recovered from seawater with an average efficiency of 98.0% at levelsof 1.3–6.6 µg/l. Analagous experiments in which arsenic (III) was used gavesimilar recoveries.

Althoughpurely thermodynamic considerations suggest that arsenic shouldexist in oxic seawaters practically entirely in the pentavalent state, equilibrium

Figure 10.1. Apparatus for the evolution and trapping of arsenic. The dimensions of theabsorption tube F are given in the inset. Source: [17]


rarely appears to be attained, probably because of the existence of biologicallymediated reduction processes. For this reason, the arsenic in most of thesewaters exists to an appreciable extent in the trivalent state, and AsIII: AsV ratiosas high as 1–1 have been found in a number of instances.

Haywood and Riley [17] found that arsenic (III) can be separated fromarsenic (V) even at levels of 2 µg/l by extracting it as the pyrrolidine dithiocar-bamate complex with chloroform. They applied this technique to samples ofseawater spiked with arsenic (V) and arsenic (III) and found that arsenic (V)could be satisfactorily determined in the presence of AsIII.

10.1.3Miscellaneous Methods

It is also possible to separate the organic arsenic compounds by column chro-matography first, and then reduce them to arsines later [16].

Organic arsenic species can be rendered reactive either by photolysis withultraviolet radiation or by oxidation with potassium permanganate or a mix-ture of nitric acid and sulfuric acids. Arsenic (V) can be determined separatelyfrom total inorganic arsenic after extracting arsenic (III) as its pyrrolidinedithiocarbamate into chloroform [15].

The nature of refractory methylarsenic compounds in estuary waters hasbeen examined by desorption chemical ionisation mass spectrometry, andmass spectrometry [18].

10.2Organocadmium Compounds

10.2.1Anodic Scanning Voltammetry

Until 1996 organocadmium compounds had not been detected in the environ-ment.

Pongratz and Hunmann [19], using differential pulse anodic scanningvoltammetry, found low levels of methyl cadmium compounds in the AtlanticOcean. Levels in the South Atlantic were approximately 700 pg/l, and thosein the North Atlantic were below the detection limit of the method, i.e., be-low 470 pg/l. It is believed that these compounds were formed as a result ofbiomethylation of inorganic cadmium.

10.3Organocopper Compounds

Mills et al. [20] carried out reversed-phase liquid chromatographic studies ofdissolved organic matter and copper–organic complexes isolated from estuar-ine waters.

10.4 Organolead Compounds 447

10.4Organolead Compounds

If we look at the intercomparison studies that have been made on the measure-ment of lead in seawater [21], two points become obvious: first, that almost noone can measure lead accurately in seawater; and second, that there is muchless lead present than anyone had estimated. With so little lead present inany form, identification of the organic moiety of an organolead compoundbecomes a major problem.

We are aided in this by our knowledge of the form in which the major an-thropogenic addition is put into the system, i.e., as the tetraethyl lead additivein gasoline. Methods have been developed for the analysis of tetraalkyl leadcompounds, both the original compound and those resulting from biologicalconversions, in the air [22]. The method for atmospheric materials uses coldtrapping (–80 ◦C) on a column of UV-1, followed by separation by gas chro-matography and analysis by atomic absorption. The method for biologicalmaterials uses benzene extraction in the presence of EDTA, followed by diges-tion to free the lead and determination by atomic absorption. While neitherof these methods is specifically designed for seawater, it should be possible toadapt them to such samples.

At least one of the organolead compounds has been shown to be a normalconstituent of the aqueous environment. Tetramethyl lead has been found asa product of anaerobic bacterial metabolism in lake sediments [23]. It shouldperhaps be sought in the air over coastal mud flats.

Bond et al. [24] examined interferences in the stripping voltammetricmethod determination of trimethyl lead in seawater using polarography andmercury 199 and lead 207 NMF. It was shown that HgII reacts with ((CH3)Pb)+ in seawater. Consequently, anodic stripping voltammetric methods fordetermining ((CH3)3Pb)2+ and inorganic PbII may be unreliable.

10.5Organomercury Compounds

Methylmercury in the marine environment may originate from industrial dis-charges or be synthesised by natural methylation processes. Fish do not them-selves methylate inorganic mercury [62, 64], but can accumulate methyl mer-cury from sea water [63]. Methylmercury has been detected in sea water onlyfrom Minamata Bay, Japan, an area with a history of gross mercury pollutionfrom industrial discharge. It has been found in some sediments but at very lowconcentrations, mainly from areas of known mercury pollution. It representsusually less than 1% of the total mercury in the sediment, and frequently lessthan 0.1% [65–67]. Microorganisms within the sediments are considered tobe responsible for the methylation [65, 68], and it has been suggested thatmethylmercury may be released by the sediments to the sea water, either in


dissolved form or attached to particulate material, and thereafter rapidly takenup by organisms [68–70].

The interest in mercury contamination, and particularly in the organicmercury compounds, is a direct reflection of the toxicity of these compoundsto man. Some idea of the proliferation of work can be derived from the re-views of Krenkel [25], Robinson and Scott [26] (460 references), and Uthe andArmstrong [27] (283 references).

All forms of mercury are potentially harmful to biota, but monomethyland dimethyl mercury are particularly neurotoxic. The lipophilic nature ofthe latter compounds allows them to be concentrated in higher trophic lev-els and the effects of this biomagnification can be catastrophic [28]. Certainspecies of microorganisms in contact with inorganic mercury produce methylmercury compounds [29]. Environmental factors influence the net amount ofmethylmercury in an ecosystem by shifting the equilibrium of the opposingmethylation and demethylation processes. Methylation is the result of mer-curic ion (Hg2+) interference with biochemical C-l transfer reactions [30].Demethylation is brought about by nonspecific hydrolytic and reductive en-zyme processes [31–33]. The biotic and abiotic influences that govern the ratesat which these processes occur are not completely understood.

Although much of the early work on the cycling of mercury pollutantshas been performed in fresh-water environments, estuaries are also subjectto anthropogenic mercury pollution [34]. A strong negative correlation existsbetween the salinity of anaerobic sediments and their ability to form methylmercury from Hg2+. As an explanation for this negative correlation the theorywas advanced that sulfide, derived by microbial reduction of sea salt sulfate,interferes with Hg2+ methylation by forming mercuric sulfide, which is notreadily methylated [35–38]. There are several reports in the literature concern-ing the methylation of Hg2+ by methylcobalamin [30, 39, 40].

Loss of Mercury on Storage

Mercury, more than most metals, is subject to loss on storage. Methylmercuryand some of the other organomercury compounds are quite volatile and easilylost fromapparently sealedbottles.Theproblemsof storagehavebeenreviewedby Mahan and Mahan [41] and by Olson [42]. Olsen felt that the greatestlosses occurred from the inorganic pool while Mahan and Mahan were moreconcerned with loss from samples with a high organic content. There have alsobeen problems associated with the kinds of bottles used for storage [43].

Stoeppler and Matthes [44] have made a detailed study of the storage be-haviour of methylmercury and mercuric chloride in seawater. They recom-mended that samples spiked with inorganic and/or methylmercury chloridebe stored in carefully cleaned glass containers acidified with hydrochloricacid to pH 2.5. Brown glass bottles were preferred. Storage of methylmercurychloride should not exceed 10 days.

10.5 Organomercury Compounds 449

From the experience of the workers in this field, it would seem that thebest method of handling seawater samples when analysing the mercury is, ifpossible, to make the analyses on shipboard as soon after collection as possible.

May et al. [45] used radiochemical studies to ascertain the behaviour ofmethylmercury chloride and mercuric chloride in seawater under differentstorage conditions. The application of 203Hg unambiguously revealed that theloss of mercury observed upon storage of unacidified seawater samples inpolyethylene bottles was due to adsorption and to the diffusion of metallic Hg(Hg0) through the container wall.

For the chemical speciation of mercury compounds the storage time andthe kind of storage are of paramount importance. After a three-day storage inbrown glass bottles 47% of 203HgCl2 added to the seawater became reducedto Hg0, but a complete reduction of HgII in the samples was not achievable.The HgII species not reduced by stannous chloride may be to a distinct ex-tent iodides and sulfides. Within 35 days of storage CH3

203HgCl decomposedinto Hg0 and HgII; 35% could be identified as Hg0 and 27% as reactive mer-cury. Strong solar radiation does not influence the transformation of 203HgII

into Hg0, but after strong three-day solar irradiation of a sample containingCH3

203HgCl, Hg0, reactive HgII, and Hg(CH3)2 could be observed. Thus cool-ing and darkness during storage are important prerequisites for the subsequentdifferentiation of mercury species. Experiments with sodium borohydride asan alternative reducing agent showed in comparison with stannous chloridea quantitative reduction of all HgII species in seawater, including CH3

203Hgwhich gave a yield of nearly 80%.

Forms of Mercury

Most of the methods of analysis for mercury actually measure inorganic mer-cury; to measure either organic or total mercury by such methods, it is nec-essary to decompose any organic mercury compounds present. This decom-position can be effected by ultraviolet irradiation of the samples. Systems ofthis sort have been described [46–48]. Since as much as 50% of the mercurymay be present in organic form [46] the differentiation between inorganic andorganic mercury can be of considerable importance.

Various workers have reported on the levels of total mercury in seawater.Generally, the levels are less than 0.2 µg/l with the exception of some parts ofthe Mediterranean where additional contributions due to man-made pollutionare found [49–52].

Fitzgerald [53] used a cold trap to concentrate mercury from large volumesof seawater.Using this technique,hecouldachieveadetection limitof 0.2 ngHg,and a coefficient of variation of 15% at the 25 ng l–1 level. Most oceanic samplescontained less than 10 ng l–1, but coastal samples could approach 50 ng l–1.

With so many variations in methodology in the literature, some sort of stan-dardisation of analytical methods is necessary. One study of standardisation


was produced by Krenkel et al. [54]. They produced a simplified extraction andclean-up procedure and recommended gas chromatography as the method ofmeasurement.

Atomic absorption spectrometry used either by direct aspiration (to deter-mine total mercury) or as an element-specific detector for gas chromatography(to determine organically bound mercury) are now discussed.


Agemian and Chau [55] have described an automated method for the deter-mination of total dissolved mercury in fresh and saline waters by ultravioletdigestion and cold vapour atomic absorption spectroscopy. A flow-throughultraviolet digester is used to carry out photo-oxidation in the automated coldvapour atomic absorption spectrometric system. This removes the chlorideinterference. Work was carried out to check the ability of the technique todegrade seven particular organomercury compounds. The precision of themethod at levels of 0.07 µg/l, 0.28 µg/l, and 0.55 µg/l Hg was ±6.0%, ±3.8%,and ±1.00%, respectively. The detection limit of the system is 0.02 µg/l.

Depending on the particular organomercury compound, recoveries werebetween 95 and 102%.

Jenne and Avotins [56] have pointed out the requirement for a strong ox-idising agent together with a strong acid for the preservation of low levels ofmercury. Both potassium permanganate and dichromate have been used as theoxidants. The former is inadequate for samples with high chloride levels sinceit would be readily consumed by chloride. Carron and Agemian [57] showedthat 1% sulfuric acid containing 0.05% potassium dichromate makes a veryeffective preservative for sub-ppb levels of mercury in water for extended pe-riods of time, especially when coupled with glass as the container. Agemianand Chau [55] preserved samples by adding 1 ml concentrated sulfuric acidand 1 ml of 5% potassium dichromate in glass containers at the start of thedilutions or sampling.

The system has a detection limit of 0.02 µg/l and is linear up to about 5 µg/l.A rate of 30 samples per hour was found to be the practical limit for the system.

Workers at the Department of the Environment UKhave described a methodfor the determination of organic and inorganic mercury in seawater [58].This method is suitable for determining dissolved inorganic mercury, andthose organomercury compounds that form dithizonates, in saline sea andestuary water. In this method inorganic mercury is extracted from the acidifiedsaline water as its dithizonate into carbon tetrachloride, but not all thesecompounds form dithizonates, and those that do not may not be determinedby this method. In general, organomercury compounds of the type RHg-X,in which X is a simple anion, form dithizonates, whereas the type R1–Hg–R2does not. Mono-methylmercury ion is extracted, although it only appears


to have a transient existence in aerobic saline water. The dithizonates aredecomposed by the addition of hydrochloric acid and sodium nitrite, andthe mercury or organomercury compound returned to the aqueous phase.Some organomercury compounds may not be completely re-extracted intothe aqueous phase. The mercury in this aqueous phase is determined by thestannous chloride reduction–atomic absorption spectroscopic technique.

Down to 4 ng/l organomercury can be determined by this method witha standard deviation of 1.3 ng/l at the zero mercury level.

Fitzgerald and Lyons [46] have described flameless atomic absorption meth-ods for determining organic mercury compounds in coastal and seawaters, us-ing ultraviolet light in the presence of nitric acid to decompose the organomer-cury compounds. In this method two sets of 100 ml samples of water are col-lected in glass bottles and then adjusted to pH 1.0 with nitric acid. One setof samples is analysed directly to give inorganically bound mercury, the otherset is photo-oxidised by means of ultraviolet radiation for the destructionof organic material and then analysed to give total mercury. The element isdetermined by a flameless atomic absorption technique, after having been col-lected on a column of 1.5% of OV-17 and 1.95% of QF1 on Chromosorb WHP(80–100 mesh). The precision of analysis is 15%. It was found that up to about50% of the mercury present in river and coastal waters is organically bound orassociated with organic matter.

Millward and Bihan [59] studied the effect of humic material on the determi-nation of mercury by flameless atomic absorption spectrometry. In both freshand seawater, association between inorganic and organic entities takes placewithin 90 min at pH values of 7 or above, and the organically bound mercurywas not detected by an analytical method designed for inorganic mercury.The amount of detectable mercury was related to the amount of humic mate-rial added to the solutions. However, total mercury could be measured afterexposure to ultraviolet radiation under strongly acid conditions.

Yamamoto et al. [60] determined picogram quantities of methyl mercuryand total mercury in seawater by gold amalgamation and atomic absorptionspectrometry. Methyl mercury was extracted with benzene and concentratedby a succession of three partitions between benzene and cysteine solution.Total mercury was extracted by wet combustion of the sample with sulfuricacid and potassium permanganate. The proportion of methyl mercury to totalmercury in the coastal seawater sampled was around 1%.

Graphite furnace atomic absorption spectrophotometry has been used forthe determination down to 5 ng/l inorganic and organic mercury in seawa-ter [61]. The method used a preliminary preconcentration of mercury usingthe ammonium pyrrolidine dithiocarbamate-chloroform system. A recoveryof 85–86% of mercury was reproducibly obtained in the first chloroform ex-tract and consequently it was possible to calibrate the method on this basis.A standard deviation of 2.6% was obtained on a seawater sample containing529 ng/l mercury.


The relative standard deviation of ten repeated determinations of 500 mldistilled water containing 10 ng mercury (II) chloride was 17.4%.


Davis et al. [71] set out to determine the concentrations of methylmercuryin sea water samples much less polluted than at Minamata Bay, namely atthe Firth of Forth, Scotland. They described a tentative bioassay methodfor determining methylmercury at the 0.06 ng/l level. Mussels from a cleanenvironment were suspended in cages at several locations in the Firth ofForth. A small number were removed periodically, homogenised, and anal-ysed for methylmercury by solvent extraction–gas chromatography, as de-scribed by Westoo [72, 73]. The rate of accumulation of methylmercury wasdetermined, and by dividing this by mussel filtration rate, the total concen-tration of methylmercury in the sea water was calculated. The methylmercuryconcentration in caged mussels increased from low levels (less than 0.01 µg/g)to 0.06–0.08 µg/g in 150 days, giving a mean uptake rate of 0.4 ng/g/d, i.e.,a 10 g mussel accumulated 4 ng/d. The average percentage of total mercuryin the form of methylmercury increased from less than 10% after 20 daysto 33% after 150 days. Davies et al. [71] calculated the total methylmercuryconcentration in the sea water as 0.06 µg/l, i.e., 0.1–0.3% of the total mer-cury concentration as opposed to less than 5–32 ng/l methyl mercury foundin Minamata Bay, Japan. These workers point out that a potentially valuableconsequence of this type of bioassay is that it may be possible to obtain es-timates of the relative abundance of methyl mercury at different sites by theexposure of “standardised” mussels as used in their experiment, in cages forcontrolled periods of time, and by comparison of the resultant accumulationsof methylmercury.

The high sensitivity and selectivity of some gas chromatographic detectorsare used to advantage in the measurement of organic mercury compounds.In the simplest approach, methyl mercury is extracted from seawater andconverted to the iodide for electron capture gas chromatography [74].

Ealy [75] alsousedconversion toalkylmercury iodides for thegas chromato-graphic determination of organomercury compounds in benzene extracts ofwater. The iodides were then determined by gas chromatograph of the benzeneextractonaglass columnpackedwith5%ofcyclohexane-succinateonAnakronABS (70–80 mesh) and operated at 200 ◦C with nitrogen (56 ml min–1) as car-rier gas and electron capture detection. Good separation of chromatographicpeaks was obtained for the mercury compounds as either chlorides, bromides,or iodides. The extraction recoveries were monitored by the use of alkylmer-cury compounds labelled with 203Hg.

An increase in specificity, permitting a lesser concern with clean-up pro-cedures, can be gained by using a microwave emission spectrometer as the


gas chromatography detector [76]. A similar method, using extraction ofmethylmercury compounds into benzene, followed by back-extraction intoL-cysteine, conversion to the chloride, back-extraction into benzene, andanalysis by electron capture gas chromatography, was devised by Chau andSaitoh [77].

The importance of selectivity and sensitivity in the detection system can beseen by examining the concentration and clean-up procedures used by Hoboet al. [78]. They used foam separation, with the addition of a surface-activecompound, to collect the organic mercury compounds, followed by extrac-tion of the foam with an organic solvent and removal of interfering heavymetals by column chromatography. Cappon and Smith [79] used a double ex-traction, in which the organic mercury compounds were first removed andthe remaining inorganic mercury was converted to methylmercury by reac-tion with tetramethyltin, and then extracted. The two organic fractions werethen cleaned up by standard procedures and determined by gas chromato-graphy.

Compeau and Bartha [80] have discussed the abiotic methylation of mer-curic ion and mercuric ion sea salt anion complexes to methylmercury bymethylcyanocobalamin in seawater and saline sediments under aerobic andanaerobic conditions.

The reaction of 30 µmol/l methylcobalamin and 60 nmol/l mercuric chlo-ride to produce methyl mercury was measured by two different methods.A spectrophotometer was used to measure changes in absorbance at 351 nmand 380 nm, respectively. These are changes characteristic of the transition ofmethylcobalamin to aquocobalamin [81].

Methyl mercury formed in the reaction was also measured gas chromato-graphically in benzene extracts of the aqueous phase [82].

An increase in absorbance at 351 nm and a concomitant decrease in ab-sorbance at 380 nm in the ultraviolet visible spectrum of methylcobalaminduring the abiotic transfer of the methyl group to Hg2+ are characteristic forthe loss of the methyl group and formation of aquocobalamin. In experimentsmonitored by both analytical techniques, gas chromatographic measurementsof methylmercury formation were in good agreement with the spectropho-tometric measurement of aquocobalamin formation from methylcobalaminat 351 nm. Aerobic versus anaerobic reaction conditions had no measurableeffect on either the methyl transfer rates, the stability of the reactants, or onthe reaction products.

A number of analytical methods for the separation of organic mercury com-pounds use an initial extraction of the organic materials with an organic sol-vent. Klisenko and Shmigidina [83] then converted both the inorganic mercuryheld in the aqueous fraction and the organic mercury in the chloroform ex-tract to dithizonate, separated the components on chromatographic columns,and determined the concentration of the various fractions by comparison withreference standards. This method is semi-quantitative at best.


10.5.3Miscellaneous

Differentiation of inorganic and organic mercury can be achieved in a numberof different ways, many of which depend upon the reduction and vapourisationof the inorganic mercury, followed by reduction [84] or oxidation [85,86] of theorganic mercury compounds, and a final measurement by atomic absorptionor mass spectrometry. Similar methods of separation of the inorganic andorganic components are used in the pretreatment of samples where the finalanalysis for mercury is to be made by neutron activation analysis [87, 88].

Sipos et al. [89] used subtractive differential pulse voltammetry at a twingold electrode to determine total mercury levels in seawater samples takenfrom the North Sea.

Ke and Thibert [90] have described a kinetic microdetermination of downto 0.05 µg/l of inorganic and organic mercury in river water and seawater.Mercury is determined by use of the iodide-catalysed reaction between CeIV

and AsIII, which is followed spectrophotometrically at 273 nm.Methylmercury has been preconcentrated from seawater by extraction with

a solution of dithizone prior to analysis by gold foil cold vapour atomic ab-sorption spectrometry [126].

Hammerschmidt and Fitzgerald [127] have studied the formation of artifactmethylmercury during extraction from a sediment reference material.

10.6Organothallium Compounds

Schedlbauer and Heumann [91] have described a sensitive analytical methodfor the determination of dimethyl thallium (Me2Tl+) in environmental samplesbypositive thermal ionisation isotopedilutionmass spectrometry (PTI-IDMS).For the necessary low detection limit and for species selective determination,PTI-IDMS was connected with species-unspecific enrichment of thallium fromocean water samples by a strongly basic anion exchanger and a species-specificextraction step, respectively. For isotope dilution, a 203Tl-enriched Me2Tl+

spike solution was synthesised and characterised with respect to its isotopiccomposition and concentration. The spike solution was stable for at least30 months under dark storage conditions at pH = 2 and 4 ◦C. The detectionlimit of the developed method is 0.4 ng/l for 500 ml ocean water samples, andthe species selectivity of Me2Tl+, compared with that of inorganic thallium, is> 500. In different surface water samples of the Atlantic Ocean from 56 ◦N to64 ◦S and in one depth profile down to 4000 m, Me2Tl+concentrations in therange < 0.4 to 3.2 ng/l were determined. This was the first time that dimethylthallium had been detected in environmental samples. Positive detection of theorganothallium compound was found in about 20% of all samples analyzed.By also analyzing the total thallium content with PTI-IDMS it was found that

10.7 Organotin Compounds 455

in these cases 3–48% of thallium exists in the methylated form. From itsoccurrence in the remote areas of the South Atlantic, from positive correlationwith the bioactivity in the corresponding ocean water, and from the fact thatMe2Tl+ is not known as an anthropogenic substance, it follows that a biogenicorigin must be assumed.

10.7Organotin Compounds

Several investigators have recently reported ng/l–µg/l concentrations of organ-otin compounds in both fresh water and marine samples. Inorganic tin,methyltins, and butyltins have been detected in marine and fresh water en-vironmental samples [92–95]. The presence of inorganic tin, butyltin, andmethyltin specieshasbeenreported inCanadian lakes, rivers, andharbours [96,97]. Both organotins and inorganic tin were reported to be highly concen-trated by factors of up to 104 in the surface microlayer relative to subsur-face water [96, 97]. Inorganic tin, mono-, di-, and trimethyltins have beendetected at ng/l levels in saline, estuarine, and fresh water samples [94, 98].Methylation of tin compounds by biotic as well as abiotic processes has beenproposed [99, 100].

Possible anthropogenic sources of organotins have recently been suggested.Both polyvinylchloride (PVC) and chlorinated polyvinylchloride (CPVC) havebeen shown to leach methyltin and dibutyltin compounds, respectively, intothe environment [101]. Monobutyltin has been measured in marine sedimentscollected in areas associated with boating and shipping. Butyltin was not de-tected in areas free of exposure to maritime activity [102]. The use of organotinantifouling coatings in particular has stimulated interest in their environmen-tal impact.

The techniques used for the investigation of organotin compounds in sea-water are atomic absorption spectrometry, gas chromatography, or gas chro-matography using AAS as detector.


Hodge et al. [92] have described an atomic absorption spectroscopic methodfor the determination of butyltin chlorides and inorganic tin in natural waters,coastal sediments, and macro algae in amounts down to 0.4 ng.

Valkirs et al. [103] determined the range of butyltin compounds in samplescollected within San Diego Bay (CA, USA). Results suggested that in certainareas of the bay the use of tin-containing antifouling paints on ships was in-creasing. Water extracted with methylisobutyl ketone and analysed by GFA-ASconsistently yielded higher tin concentrations than the same samples analysed


by a hydride reduction/flame atomic absorption method [92, 98], suggestinga non-hydride reducible tin fraction was present. Gas chromatography–massspectrometry confirmed the presence of tributyltin in an environmental ma-rine water sample following derivatisation to the hydride species.

Surface and bottom water samples were collected in 500 ml, 1, or 4 litrepolycarbonate bottles. Polycarbonate bottles have been shown to retain 97%of an initial spike of bis(tri-n-butyltin) oxide in seawater at a concentrationof 0.5 mg/l over a weeklong period [104]. Samples were analysed immediatelyafter collection and transported to the laboratory, or were stored frozen at–20 ◦C and analysed at a later date. Frozen storage has been shown to be ef-fective in preserving sample stability with respect to monobutyltin, dibutyltin,and tributyltin concentrations for a period of at least 100 days.

An estimate of the accuracy of both analytical methods was performed onbis(tri-n-butyltin) oxide and tri-n-butyltin chloride solutions (8.9–35.6 µg/l)prepared in filtered (0.45 µm) near-shore seawater free of detectable organ-otins. Average mean recoveries of 92.8% by both methods were determined fortributyltin standard solutions. Low ng/l levels of mono-, di-, and tributyltinwere found in samples taken from San Diego Bay.

Valkirs et al. [105] have conducted an interlaboratory comparison or deter-minations of di- and tributyltin species in marine and estuarine waters usingtwo methods, namely hydride generation with atomic absorption detectionand gas chromatography with flame photometric detection. Good agreementwas obtained between the results of the two methods. Studies on the effect ofstoring frozen samples prior to analysis showed that samples could be storedin polycarbonate containers at – 20 ◦C for 2–3 months without significant lossof tributyltin.

Burns et al. [106] used electrothermal AAS to determine inorganic andbutyltin in seawater. The butyltin is extracted into toluene and the inorganictin extracted as its Sn(IV) 8-hydroxyquinoline chelate into chloroform. Thedetection limit was 0.7 ng of tin.

The separation of mono-, di-, and tributyltin species in seawater by isocraticion exchange liquid chromatography coupled to hydride generation AAS hasbeen reported by Schulze and Lehmann [107]. Reported detection limits are31, 40, and 27 mol/l, respectively.


Studies by Braman and Tompkins [98] have shown that nonvolatile methyltinspeciesMenSn(4–n)+

aq (n = 1–3), areubiquitous atng/l concentrations innaturalwaters including both marine and fresh water sources. Their work, however,failed to establish whether tetramethyltin was present in natural waters becauseof the inabilityof themethodsused to trap this compoundeffectivelyduring the


combined preconcentration purge and reductive derivatisation steps employedto generate volatile organotin hydrides necessary for tin specific detection.

Tin compounds are converted to the corresponding volatile hydride (SnH4,CH3SnH3, (CH3)2SnH2, and (CH3)3SnH) by reaction with sodium borohydrideat pH 6.5 followed by separation of the hydrides and then atomic absorptionspectroscopyusingahydrogen-richhydrogen–airflameemission typedetector(Sn–H band).

The technique described has a detection limit of 0.01 ng as tin, and henceparts per trillion of organotin species can be determined in water samples.

Braman and Tompkins [98] found that stannane (SnH4) and methylstan-nanes (CH3SnH3, (CH3)2SnH2 and(CH3)3SnH)couldbe separatedverywell onacolumncomprisingsiliconeoilOV-3 (20%w/w)supportedonChromosorbW.A typical separation achieved on a water sample is shown in Fig. 10.2.

Thegenerationandrecoveryof stannane,methyl-, dimethyl-, and trimethyl-stannane were studied in seawater. Average tin recoveries for the six samplesanalysed, to which were added 0.4–1.6 µg methyltin compounds and 3 nginorganic tin, ranged from 96 to 109%. Reanalysis of analysed samples showsthat all methyltin and inorganic tin is removed in one analysis procedure.

A number of natural waters, from in and around the area of Tampa Bay,Florida, were analysed for tin content. All samples were analysed withoutpretreatment. Samples that were not analysed immediately were frozen untilanalysis was possible. Polyethylene bottles, 500 ml volume, were used for sam-ple acquisition and storage. The results of the analyses appear in Table 10.1.The average total tin content of fresh, saline, and estuarine waters are 9.1, 4.2,and 12 ng/l, respectively. Approximately 17–60% of the total tin present wasfound to be in the methylated forms. The saline waters appear to have thehighest percentage of methylated tin compounds; 60% of the total tin present

Figure 10.2. Environmental sample analysis of organotin compounds. a Environmentalsample, Tampa Bay. b Typical blank. Source: [98]


Table 10.1. Analysis of saline and estuarine water samples. Data are average of duplicates(given in ng/l; results in parentheses are percentages of total tin) [98]

Sample Tin(IV) Methyl tin Dimethyl tin Trimethyl tin Total tin

Gulf of Mexico,Sarasota

62 (73) 15 (18) 7.0 (8.3) 0.98 (1.2) 85

Gulf of Mexico,Fort De Soto

2.2 (6.0) ND 0.74 (2.0) 0.71 (20) 3.6

Gulf of Mexico,St. Petersburg

4.5 (54) 0.62 (7.4) 3.2 (39) ND 8.3

Old Tampa Bay,Oldsmar

0.3 (9.7) 0.86 (33) 0.88 (34) 0.61 (24) 2.6

was found to be in the methylated forms, and the dimethyltin form contributesapproximately half of this value.

The above procedure, although valuable in itself, is incomplete in thatany monobutyltin species present escape detection. Excellent recoveries ofmonobutyltin species are achieved with tropolone.

Meinima et al. [108] studied the effect of combinations of various sol-vents with 0.05% tropolone on the recoveries of mono-, di-, and tributyltinspecies either individually or simultaneously present in aqueous solutions.The results obtained by gas chromatography–mass spectrometry after methy-lation show that Bu3Sn and Bu2Sn recoveries appear to be almost quantita-tive both for neutral and from hydrobromic acid-acidified aqueous solutions.Butyltin recovery appears to be influenced by the presence of hydrobromicacid in that, in general, recovery rates are higher from solutions acidifiedwith hydrobromic acid than from non-acidified solutions. Bu3Sn and Bu2Snrecoveries remain fairly constant with ageing of an aqueous solution of thesespecies over a period of several weeks. Butyltin recoveries, however, do de-crease with time to a notable extent (20–40%), most likely as a result ofadsorption/deposition of Bu1Sn species to the glass wall of the vessel. Addi-tion of hydrobromic acid obviously affects the desorption of these species,as recovery of Bu1Sn species increases to almost the same values as ob-tained from hydrobromic acid-acidified freshly prepared aqueous solutionsof species.

Jackson et al. [109] devised trace speciation methods capable of ensuringdetection of tin species, along with appropriate preconcentration and derivati-sation without loss, decomposition, or alteration of their basic molecular fea-tures.

They describe the development of a system employing a Tenax GC filledpurge and trap sampler, which collects and concentrates volatile organotinsfrom water samples (and species volatilised by hydrodisation with sodiumborohydride), coupled automatically to a gas chromatograph equipped with


a commercial flame photometric detector modified for tin-specific detec-tion [110–112].

The system was applied to the analysis of a series of water samples obtainedfrom the Chesapeake Bay at both industrially polluted and relatively pristinesites.

The volatile species were determined by the purge and trap gas chromato-graphic flame photometric detector method without the sodium borohydridereduction step. Non-volatile species were volatilised, trapped, and detectedwith the volatile species present in the samples. Generally no changes inthe concentration of the volatile species were observed with the hydridereduction step. An increase in the amount detected, as with Me2SnH2 andBuSnH3, indicated that the solvated cations were also present in the watersamples along with the respective free stannane. The prevailing gas chro-matographic peak at 1.75 ± 0.04 min has not been completely resolved oridentified. The retention time is within the range of both dimethyl sulfide(Me2S) and Me2SnH2. Dimethyl sulfide is known to be a microbial metabo-lite present in environmental water; therefore, this is a possible interfer-ence.

Brinckmann and co-workers [113] used a gas chromatographic methodwith or without hydride derivatisation for determining volatile organotincompounds (e.g., tetramethyltin) in seawater. For nonvolatile organotin com-pounds a direct liquid chromatographic method was used. This system em-ploys a “Tenax GC” polymeric sorbent in an automatic purge and trap (P/T)sampler coupled to a conventional glass column gas chromatograph equippedwith a flame photometric detector. Flame conditions in the FPD were tunedto permit maximum response to SnH emission in an H-rich plasma, as de-tected through narrow bandpass interference filters (610 ± 5 nm). Two modesof analysis were used:

(1) Volatile stannanes were trapped directly from sparged 10–50 ml watersamples with no pretreatment.

(2) Volatilised tin species were trapped from the same or replicate water sam-ples following rapid injection of aqueous, excess sodium borohydride so-lution directly into the P/T sparging vessel immediately prior to beginningthe P/T cycle.

10.7.3Hydride Generation Gas Chromatography–Microwave Induced Atomic Emission Spectrometry (HGGC–MIAES)

A recent method [114] for the speciation of organotin compounds in seawatercombines solid-phase extraction, online HG, and GC with MIAE. This methodenjoys a 0.5 pg detection limit for tin.


10.7.4Thermal Desorption-Gas Chromatography–Inductively Coupled Plasma Mass Spectrometry (TDGC–ICPMS)

Vercauteren et al. [115] have investigated the extraction and preconcentra-tion capabilities of a new extraction technique in which stir bar sorptiveextraction was combined with the separation power of capillary gas chro-matography and the low limits of detection of ICPMS for the determinationof the organotin compounds tributyltin and triphenyltin in aqueous standardsolutions, harbour water, and mussels after digestion with tetramethylammo-nium hydroxide. Throughout, tripropyltin for tributyltin and tricyclohexyltinfor triphenyltin were used as internal standards to correct for variations inthe derivatisation and extraction efficiency. Calibration was accomplished bymeans of single standard addition. Derivatisation to transform the trisubsti-tuted compounds into sufficiently volatile compounds was carried out withsodium tetraethylborate. The compounds were extracted from their aqueousmatrix using a stir bar of 1 cm length, coated with 55 µl of poly (dimethyl-siloxane). After 15 min of extraction, the stir bar was desorbed in a thermaldesorption unit at 290 ◦C for 15 min, during which the compounds were cold-trapped on a precolumn at –40 ◦C. Flash heating was used to rapidly transferthe compounds to the gas chromatograph where they were separated on a cap-illary column with a poly(dimethylsiloxane) coating. After separation, thecompounds were transported to the ICP by means of a homemade heated(270 ◦C) transfer line. Monitoring of the 120Sn+ signal by ICPMS during therun of the gas chromatograph provided extremely lower limits of detectionin water: 0.1 pg/l (procedure) and 10 fg/l (instrumental) and a repeatability of12% RSD (n = 10). In harbour water, concentrations were found of 200 pg/lfor tributyltin. The accuracy of the method was checked by the determinationfor triphenyltin in CRM 477 (mussel tissue) and comparison of the results tothat of an analysis of the same material with a classical liquid–liquid extractionwith isooctane.

Encinar et al. [128] used a spike containing 119Sn enriched mono-, di-, andtributyltin to determine butyltin compounds in seawater GC–ICPMS. Reversespiking was used to assess species transformation during derivativisation.

To determine organotin compounds in amounts down to 1–5 ng/l, McAvoyet al. [129] ethylated the seawater, isolated organotin solid-phase micro extrac-tion and determined it by micro capillary GCMS.

Pocurull et al. [130] used the online solid-phase extraction GCMS precon-centrationmethod to determine organotin antifouling compounds in seawater;10 ng/l detection limits were achieved using a 10 ml sample.

Ferrer and Barcelo [131] used online solid-phase–liquid chromatographymass spectrometry for the simultanous determination of organotin antifoulingherbicides in marine water. The solid-phase extraction was carried out onpolymeric cartridges after percolation of 100 ml of the seawater sample, and


showed recoveries of 96–111% of the antifouling compound. Detection limitswere about 5 ng/l.

Tao et al. [132] achieved a 0.01 pg/l detection limit for organotin species inseawater using GC–ICPMS. This method involved derivitisation using a Grig-nard reagent, sodium tetraethylborate preconcentration by extraction intohexane, a programmed temperature vaporisation, and the operation of a shieldtorch at normal conditions. The use of the programmed temperature vapori-sation enabled the injection of large sample volumes, up to 100 µl without lossof analyte.


Ebdon and Alonso [116] have determined tributyltin ions in estuarine watersby high-performance liquid chromatography and fluorometric detection usingmorin as a micellar solution. Tributyltin ions were quantitatively retained from100–500 mL of sample on a 4 cm-long ODS column. After washing off the saltswith 20 ml of distilled water the ODS column was back-flushed with methanol-water (80 + 20) containing 0.15 M ammonium acetate, on a 25 cm-long PartisilSCX analytical column. The eluant from the column (1 ml min–1) was mixedwith fluorometric reagent (acetic acid, (0.01 M); morin (0.0025% m/v); Tri-ton X-100, (0.7% m/v 2.5 ml/min)) for detection at 524 nm with excitation at408 nm. The detection limit (2σ) is 16 ng of tributyltin (as Sn).

10.7.6Miscellaneous

Smith [117] discussed the determination of tin in water. In the determinationof low concentrations of the order of 40 ng of trialkyltin chlorides in sea waterit has been observed that these compounds are very volatile and are easily lostupon evaporation with acid. Quantitative recovery of tin is, however, obtainedin the absence of chloride ion during evaporation with acid. Preliminary re-moval of chlorides from sea water by passage down a column of IRA 400resin before digestion with acid completely overcame loss of tin on subsequentevaporation, with acid giving a tin recovery of 90%.

Duhamel et al. [118] investigated the behaviour of this (tributyltin) ox-ide and tributyltin chloride in saline water. The effects of salinity, pH, light,and oxygen were investigated. Debutylation due to the formation of insolublecompounds occurred under saline conditions.

Kenis and Zirino [119] determined tri-n-butyltin oxide directly in seawaterat microgram per litre levels at a hanging drop mercury electrode, and ata mercury film rotating-disk glassy electrode by differential pulse anodic-stripping voltammetry. The hanging drop mercury electrode responded to tri-n-butyltin oxide additions in sea water purged either with nitrogen (pH 8.2)


or carbon dioxide (pH 4.8) with two distinct stripping peaks. The mercuryfilm rotating-disk glassy electrode produced stripping peaks due to tri-n-butyltin oxide at pH 8.2, and no response at pH 4.8. Peak potentials and peakheights varied depending upon the history of the water samples, suggesting aninfluence by dissolved basic materials. The limit of detection for tri-n-butyltinoxide in sea water on the hanging drop electrode was 5 ng/l. The hanging dropelectrode responded to additions of inorganic tin in sea water at pH 4.8 and8.2, whereas the rotating-disk electrode did not respond at either pH.

Luskima and Syavtsillo [120] have described a spectrophotometric proce-dure utilising phenylfluorone for the determination of organotin compoundsin water.

A method for the determination of organotin compounds by anodic strip-ping polarography has been published [121]. It has yet to be applied to sea-water. Since the sensitivity permits the measurement of 0.01 ppm of the tincompounds, it is likely to be not quite sensitive enough for seawater and wouldrequire a preconcentration step.

Laughlin et al. [122] analysed chloroform extracts of tributyltin dissolvedin seawater using nuclear magnetic resonance spectroscopy. It was shown thatan equilibrium mixture occurs which contains tributyltin chloride, tributyl tinhydroxide, the aquo complex, and a tributyltin carbonate species. Fluorometryhas been used to determine triphenyltin compounds in seawater [123]. Triph-enyltin compounds in water at concentrations of 0.004–2 pmg/l are readilyextracted into toluene and can be determined by spectrofluorometric mea-surements of the triphenyltin-3-hydroxyflavone complex.

Tri-, di-, and monobutyl, and di- and monoethyltin compounds, did notfluoresce under the conditions used for the determination of triphenyltin.However, trimethyltin compounds react in a similar manner with 3-hydroxyl-flavone, and although the emission maximum is at approximately 510 nm,this is not sufficiently different from the emission maximum of triphenyltincompounds (approximately 495 nm) for these compounds to be determined inthe presence of one another.

Spiking recoveries by the above procedure carried out on standard solutionsof triphenyltin chloride in various types of water ranged from 74% at the 4 µg/ltin level (relative sd 8.9%) to 93.6% at the 2 mg/l tin level (relative sd 4.2%).

Organotincompoundshavepreconcentratedoncationexchangeresins [124]and XAD-2 resin [125] (See Table 9.3).

References

1. Braman RS, Foreback GC (1973) Science 182:12472. Woolson EA (1975) Arsenical pesticides. ACS symposium series no. 7. American

Chemical Society, Washington, DC3. Penrose WR (1974) Crit Rev Environ Contr 4:4654. Andraea MO (1977) Anal Chem 49:820

References 463

5. Braman RS, Johnson DL, Foreback CC, Ammons JM, Bricker JL (1977) Anal Chem49:621

6. Howard AG, Arbab-Zavar MH (1981) Analyst 106:2137. Arbab-Zavar MH, Howard AG (1980) Analyst 105:7448. Hinners TA (1980) Analyst 105:7519. Pierce FD, Brown HR (1977) Anal Chem 49:1417

10. Talmi Y, Bostick DT (1975) Anal Chem 47:214511. Edmonds JS, Francesconi KA (1976) Anal Chem 48:201912. Johnson DL, Braman RS (1975) Deep-Sea Res 22:50313. Braman RS, Johnson DL, Foreback CC, Ammons JM, Bricker JL (1977) Anal Chem

49:62114. Persson J, Irgum K (1982) Anal Chim Acta 138:11115. Lundgren G, Lundwork L, Johansson G (1974) Anal Chem 46:102816. Yamamoto M (1975) Soil Science Society of America Proceedings 39:85917. Haywood MG, Riley JP (1976) Anal Chim Acta 85:21918. De Bettencourt Am, Florencio MH, Duarte MFN, Gomes MLR, Boas LFCV (1994) Appl

Organometall Chem 8:4319. Pongratz R, Heumann KG (1996) Anal Chem 68:126220. Mills GL, McFadden E, Quinn JG (1987) Mar Chem 20:31321. Patterson C, Settle D (1976) Mar Chem 4:38922. Chau YK, Wong PTS, Saitoh H (1976) J Chromatogr Sci 14:16223. Wong PTS, Chau YK, Luxon PL (1975) Nature 253:26324. Bond AM, Bradbury JR, Hanna PJ, Howell GN, Hudson HA, Strother S, O’Connor MJ

(1984) Anal Chem 56:239225. Krenkel PA (1973) Int Crit Rev Environ Cont 3:30326. Robinson S, Scott WB (1974) A selected bibliography on mercury in the environment,

with subject listing. Life science miscellaneous publication. Royal Ontario Museum,p 54

27. Uthe JF, Armstrong FAJ (1974) Toxicol Environ Chem Rev 2:4528. D’ltri PA, D’ltri PM (1978) Environmental Management 2:329. Jensen S, Jernelov A (1969) Nature 753:22330. Wood JM (1974) Science 183:104931. Furakawa K, Tonomura K (1971) Agr Biol Chem 35:60432. Furakawa K, Tonomura K (1972) Agr Biol Chem 36:21733. Furakawa K, Tonomura K (1972) Agr Biol Chem 36:244134. Brinckman FE, Iverson WP (1975) In: Church T (ed) Marine chemistry in the coastal

environment. American Chemical Society symposium 18. American Chemical Society,Washington, DC

35. Fagerstrom T, Jernelov A (1971) Water Res 5:12136. Yamada M, Tonomura K (1972) J Ferment Tech 50:15937. Yamada M, Tonomura K (1972) J Ferment Tech 50:89338. Yamada M, Tonomura K (1972) J Ferment Tech 50:90139. Bertilsson L, Neujahr HY (1971) Biochemistry 10:280540. Imura NE, Sukegawa SK, Pan K, Nagao K, Kim JK, Kwan T, Ukita T (1971) Science

172:124841. Mahan KI, Mahan SE (1977) Anal Chem 49:66242. Olson KR (1977) Anal Chem 49:2343. Heiden RW, Aikens DA (1977) Anal Chem 49:66844. Stoeppler M, Matthes W (1978) Anal Chim Acta 98:38945. May K, Reisinger K, Rlucht R, Stoeppler M (1980) Sendeedenck vom Wasser 55:63


46. Fitzgerald WF, Lyons WB (1973) Nature 242:45247. Frimmel F, Winkler HA (1975) Zeitung Wasser-Abwasser-Forschung 8:6748. Kiemeneij AM, Kloosterboer JG (1976) Anal Chem 48:57549. Thibaud Y (1971) Science et Peche, Bulletin Instrumental Peche Maritime 209:150. Cumont G, Viallex G, Lelievre H, Babenricht P (1972) Revue International Oceanogra-

phie Mediterranean 26:9551. Renzoni A, Baldi F (1975) Accua and Aria 59752. Stoeppler M, Backhaus F, Matthes W et al. (1976) Proc. Verb. XXVth Congress and

Plenary Assembly of ICSEM, Split53. Fitzgerald WF (1975) Mercury analyses in seawater using cold trap preconcentration

and gas-phase detection. In: Gibb RP Jr (ed), Analytical Methods in Oceanography,American Chemical Society, Washington, DC, pp 99–109, 147

54. Krenkel PA, Burrow W, Dickinson W (1975) Standardization of methylmercury anal-ysis. Report COM-75-10673

55. Agemian H, Chau ASY (1978) Anal Chem 50:1356. Jenne EA, Avotins P (1975) J Environ Qual 4:42757. Carron J, Agemian H (1977) Anal Chim Acta 92:6158. Department of the Environment and National Water Council (UK) (1985) Mercury in

waters, effluents, soils, and sediments, etc., additional methods. Pt 22-AG ENV. HMSO,London

59. Millward GE, Bihan AI (1978) Water Res 12:97960. Yamamoto J, Kaneda Y, Hisaka Y (1983) Int J Environ Anal Chem 16:161. Filipelli M (1984) Analyst 109:51562. Penncicchioni A, Marchetti R, Gaggino GF (1976) J Environ Qual 5:45163. Pentreath RJ (1976) J Exp Mar Biol Ecol 25:5164. Pentreath RJ (1976) J Exp Mar Biol Ecol 25:10365. Olsen BH, Cooper RC (1976) Water Res 10:11366. Andrer AW, Harris RC (1973) Nature 245:25667. Bartlett PO, Craig PJ, Morton SF (1977) Nature 267:60668. Shin E, Krenkel PA (1976) J Water Pollut Cont Fed 48:47369. Jeruelev A (1970) Limnol Oceanogr 15:95870. Langley DG (1973) J Water Pollut Cont Fed 49:4471. Davis IM, Graham WC, Pirie JM (1979) Mar Chem 7:1172. Westhoo G (1966) Acta Chim Sci Anal 20:213173. Westhoo G, Johanssen B, Ryhage R (1978) Acta Chem Scand 28:227774. Longbottom JE, Dressman RC, Lichtenberg JJ (1973) J Assoc Off Anal Chem 56:129775. Ealy J, Schulz WD, Dean DA (1974) Anal Chim Acta 64:23576. Talmi U (1975) Anal Chim Acta 74:10777. Chau YK, Saitoh H (1973) Int J Environ Anal Chem 3:13378. Hobo T, Ogura T, Suzuki S, Araki S (1975) Bunseki Kagaku 24:28879. Cappon CJ, Smith JC (1977) Anal Chem 49:36580. Compeau G, Bartha R (1983) Bull Environ Contam Toxicol 31:48681. Dolphin D (1971) In: McCormick DB, Wright LD (eds) Methods in enzymology, vol.

18, part c. Academic Press, New York82. Blum J, Bartha R (1980) Bull Environ Contam Toxicol 25:40483. Klisenko MA, Shmigidina AM (1974) Gig Sahit 1974:6484. Kamada T, Hayashi Y, Kumamaru T, Yamamoto Y (1973) Bunseki Kagaku 22:148185. Baltisberger RJ, Knudson CL (1974) Anal Chim Acta 73:26586. El-Awady AA, Miller RB, Clark MJ (1976) Anal Chem 48:11087. Van der Sloot HA, Das HA (1974) Anal Chim Acta 73:235

References 465

88. De Jong EG, Wiles DR (1976) Journal of the Fisheries Research Board of Canada 33:132489. Sipos L, Nurenberg HW, Valenta P, Branica M (1980) Anal Chim Acta 115:2590. Ke PJ, Thibert RT (1973) Mikrochim Acta 3:41791. Schedlbauer OF, Heumann KG (1999) Anal Chem 71:545992. Hodge VF, Seidel SL, Goldberg ED (1979) Anal Chem 51:125693. Waldock MJ, Miller D (1983) ICES paper CM 1983 E:12 (mimeograph). International

Council for the Exploration of the Sea, Copenhagen94. Tugrul S, Balkas TL, Goldberg ED (1983) Mar Pollut Bull 14:29795. Müller MD (1984) Fresen Z Anal Chem 317:3296. Maguire RJ, Chau YK, Bengert GA, Hale EJ, Wong PTS, Kramar O (1982) Environ Sci

Technol 16:69897. Maguire RJ (1984) Environ Sci Technol 18:29198. Braman RS, Tompkins MA (1979) Anal Chem 51:1299. Ridley WP, Dizikes LJ, Wood JM (1977) Science 197:329

100. Guard HE, Cobet Ab, Coleman WM (1981) Science 213:710101. Boettner EA, Ball GL, Hollingsworth Z, Aquino R (1982) Report OH EPA-600/SL-

81-602. Health Effects Research Laboratory, US Environmental Protection Agency,Cincinatti, OH

102. Seidel LS, Hodge VF, Goldberg ED (l980) Thalassia Jugoslavia 16:209103. Valkirs AO, Seligman PF, Stang PM, Homer V, Lieberman SH, Vafa G, Dooley CA (1986)

Mar Pollut Bull 17:319104. Dooley CA, Homer V (1983) Organotin compounds in the marine environment: uptake

and sorption behaviour. Technical report no. 917. Naval Ocean Systems Center, SanDiego, CA, USA

105. Valkirs AO, Seligman PF, Olsen GJ, Brinckman FE, Matthias CL, Bellama JM (1987)Analyst 112:17

106. Burns DT, Harriott M, Glockling F (1987) Fresen Z Anal Chem 327:701107. Schulze G, Lehmann C (1994) Anal Chem 288:215108. Meinema HA, Burger N, Wiersina T (1978) Environ Sci Technol 12:288109. Jackson JA, Blair WR, Brinckman FE, Iveson WP (1982) Environ Sci Technol 16:111110. Aue WA, Flinn GC (1977) J Chromatogr 142:145111. Huey C, Brinckman FE, Grim S, Iveson WP (1974) In: Freitas ASW, Kushner DJ, Quadri

DSU (eds) Proceedings of the international conference on the transport of persistentchemicals in aquatic ecosystems. National Research Council of Canada, Ottawa, ppII-73–II-78

112. Nelson JD, Blair W, Brinckman FE et al. (1973) Appl Microbiol 26:321113. Brinckmann FE (1981) In: Wong CS et al. (eds) Proceedings of a NATO advanced

research institute on trace metals in seawater, Sicily, Italy. Plenum Press, New York114. Dowling TM, Uden PC (1993) J Chromatogr 644:153115. Vercauteren J, Peres C, Devos C, Sandra P, Vanhaeke F, Moens L (2001) Anal Chem

73:1509116. Ebdon L, Alonso G (1987) Analyst 112:1551117. Smith JD (1970) Nature 225:103118. Duhamel K, Blanchard G, Dorange G, Martin G (1987) Appl Organometall Chem

1:133119. Kenis A, Zirino A (1983) Anal Chim Acta 149:157120. Luskina BM, Syavtsillo SV (1969) Nov Obl Prom-Sanit Khim 186121. Waggon H, Jehle D (1975) Nahrung 19:271122. Laughlin RB, Guard HE, Coleman, WM (1986) Environ Sci Technol 20:201123. Bluden SJ, Chapman AH (1978) Analyst 103:1266


124. Ochsenkuhn-Petropolu M, Ochsenkuhn KM, Parissakis G, Glannakis An, Smith HD(1995) Canadian J Appl Spectros 40:66

125. Nagase T, Tao H, Imagawa T, Miyazaki A (1992) Shigei to Konkyo 1:263126. Schintu M, Kauri T, Contu A, Kudo A (1987) Ecotox Environ Safe 14:208127. Hammerschmidt CR, Fitgerald WF (2001) Anal Chem 73:5930128. Encinar JR, Villar MIM, Santamaría VG, Alonso JIG, Sanz-Medel A (2001) Anal Chem

73:3174129. McAvoy DC, Schatowitz B, Jacob M, Hauk A, Eckhoff WS (2002) Environ Toxicol Chem

21:1323130. Pocurull E, Brossa L, Borrull F, Marce RM (2000) J Chromatogr A 885:361131. Ferrer I, Barcelo D (1999) J Chromatogr A 854:197132. Tao H, Rajendran R, Quetel CR, Nakazato T, Tominaga M, Miyazaki A (1999) Anal

Chem 71:4208

11 Elemental Analysis

11.1Boron

Ball et al. [1] have described a method for determining down to 20µg/l boron inestuarine waters. A dc argon plasma emission spectrometer was used. Quench-ing of the plasma by high solute concentrations of easily ionised elements suchas alkali metals, as well as high and variable electron density, may be avoidedby dilution. The method was found to be more sensitive, equally precise,less subject to interference and with a wider linear analytical range than thecarmine spectrophotometric method. Very few interferences were noted whenthis technique was tested. There is a minor interference from the differentialenhancement of tungsten relative to boron in solutions containing high con-centrations of alkali metals. The effect of this is to increase the backgroundwhen estuary water is being analysed, and it can be mitigated by using syn-thetic estuary water as a blank, by dilution, or by analysis by the method ofstandard additions.

An oscillopolarographic method gave a detection limit of 3 ×10–9 mol/dm–3

and a linear range of 3 ×10–9 to 7 ×10–7 mol/dm–3 in the determination ofboron in seawater.

Traces of boron in seawater have been determined by flow injection analysiswith spectrophotometric detection at 415 nm using G 30 methine H. The linearrange was 1–10 mg/l boron with a detection limit of 0.017 mg/l [3].

11.2Total Iodine

Schnepfe [4] has described a method for the determination of total iodineand iodate in seawater. One per cent aqueous sulfamic acid (1 ml) is addedto seawater (10 ml), then it is filtered, if necessary, and the pH adjusted to 2.After 15 min, 1 ml 0.1 M sodium hydroxide and 0.5 ml 0.1 M potassium per-manganate are added and the steam bath heated for 1 h. The cooled solutionis filtered, the residue washed, the filtrate plus washings is diluted to 16 mland 1 ml of a 0.25 M phosphate solution (containing 0.3 µg iodine as IO–

3 per

468 11 Elemental Analysis

ml) added at 0 ◦C, then 0.7 ml 0.1 M Fe11 in 0.2% (v/v) sulfuric acid, 5 ml 10%aqueous sulfuric acid–phosphoric acid (1:1) at 0 ◦C, and 2 ml starch-cadmiumiodide reagent are added. The solution is diluted to 25 ml and after 10–15 min-utes the extinction of the starch–iodine complex is measured in a 5 cm cell.For iodate the procedure is as described above except that the oxidation stagewith sodium hydroxide–potassium permanganate is omitted and only 0.2 mlFe solution is used; iodate standards were used in both procedures. The totaliodine procedure is relatively free from interference by foreign ions but theiodate procedure is affected by bromate and sulfite. Down to 0.1 µg iodine canbe determined in the presence of 500 mg chloride and 5 mg bromide by thisprocedure.

The classical method for the determination of iodide in seawater was de-scribed by Sugawara [5]. Various workers [6, 7] have improved the originalmethod. Matthews and Riley [6] modified the method by concentrating iodideby means of coprecipitation with chloride using silver nitrate (0.23 g per 500 mlseawater). Treatment of the precipitate with aqueous bromine and ultrasonicagitation promote recovery of iodide as iodate which [15] when reacted withexcess of iodide ions under acid conditions, yields I–

3, which are determinedeither spectrophotometrically or by photometric titration with sodium thio-sulfate. Photometric titration gave a recovery of 99.0 ± 0.4% and a coefficientof variation of ±0.4% compared with 98.5 ± 0.6% and ±0.8%, respectively, forthe spectrophotometric procedure.

Tsunogai [7] carried out a similar coprecipitation allowing a 20-hour stand-ing period to ensure that iodide is fully recovered in the silver chloride copre-cipitate. Again, the iodide is oxidised to iodate prior to spectrophotometricdetermination of the latter. This procedure also includes a step designed toprevent interference by bromine compounds.

Kesari et al. [8] have recently described a sensitive spectrophotometricmethod for determining iodine in seawater.

11.3Organic Nitrogen

The quantification of nitrogen in its soluble organic forms has interested ma-rine chemists engaged in hydrographic, primary production and pollutionstudies [20, 21]. The concentrations of particular dissolved organic nitro-gen species in seawater may be low, and while methods have been devel-oped for some specific compounds (notably the amino acids) the bulk ofthe dissolved organic nitrogen remains uncharacterised. Whilst a large por-tion of the dissolved organic nitrogen in seawater is probably derived fromsoluble exudates from algae, it will also contain a wide range of other com-pounds including animal proteins, urea, and other excretory products, nucleicacids, etc.

11.3 Organic Nitrogen 469

Dissolved Organic Nitrogen is Commonly Determined in Four Stages:

1. Removal of particulate organic nitrogen by filtering through a 0.45 µm poremembrane or equivalent glass-fibre filter

2. Oxidation of soluble organic nitrogen to inorganic forms of nitrogen3. Determination of total inorganic nitrogen4. Correction for inorganic nitrogen species present before oxidation.

The micro Kjeldahl method involves reducing all of the organic nitrogen toammonia, then distilling the ammonia into an absorbing solution and deter-mining it colorimetrically [11, 12, 22].

A comparative study of the variations in final measurements of ammoniahas been carried out by Astrani [9]; a Russian version [10] could also be usedfor the measurement of organic nitrogen in particulate matter.

Mertens et al. [11] and Stevens [12] designed semiautomated versions of themicro Kjeldahl which avoided the distillation step altogether. In their versions,after the digestion step the digestion solution was diluted and the ammoniadetermined with an ammonia probe. The limitation on the sensitivity, then,is the sensitivity of the ammonia probe. This limits the method to the moreproductive oceanic waters.

Another approach to the organic nitrogen problem is to use persulfate wetoxidation toconvert thenitrogen tonitrateornitrite, inplaceof the reduction toammonia [13,14,24,25]. Results are fully comparable with those from the microKjeldahl digestion but the technique is far simpler. The precision should also behigher, since the final step in the measurement, the colorimetric determinationof nitrite, is much more precise than any of the ammonia methods.

The Adamski [13] procedure is semiautomated. It gives an accuracy of±8.1% and a precision of ±8.2% for seawater samples spiked with 3–35 µg/lorganic nitrogen. This procedure is based on the indophenol blue methodand was employed using a Technicon AutoAnalyzer II system with the appro-priate accessories. Various workers have described automated procedures fordetermination of low levels of organic nitrogen in seawater [15, 16].

Ultraviolet photo-oxidation techniques can be used as a method for organicnitrogen. The organic nitrogen compounds are oxidised to nitrate and nitritethen determined by the standard seawater analysis methods [17, 18, 23].

The oxidation of urea was not complete. The incomplete step in the photol-ysis of urea is much more likely to be hydrolysis of urea to produce ammonia,since the oxidation of ammonia is known to occur under the conditions of theexperiment.

If the freeze drying method for total organic carbon devised by Gordonand Sutcliffe [29] is used with the final determination of carbon being run ona commercial CHN analyser, the analysis of total nitrogen is also obtained.If analyses for the inorganic nitrogen compounds have also been run on thesample, organic nitrogen can be calculated by difference.


Of the methods in common use, the two wet oxidation methods offer thebest possibilities for further development. The photo-oxidation method inparticular is well suited to automatic analysis. In the version of the total organiccarbon method published by Collins and Williams [30] the effluent from thequartz photolysis coil could as easily be diverted to the nitrate analysis unit;if inorganic nitrogen were also measured, organic nitrogen could becomea routine automatic method.

Shepherd and Davies [26] have described a semiautomated version of thealkaline peroxodisulfate procedure for the determination of total dissolvedorganicnitrogen in seawater. Theseworkers carriedout experimentsona rangeof contaminated and clean seawaters using a version [27] of Koroleff ’s [24]alkaline peroxodisulfate oxidation technique.

Shepherd and Davies [26] made some modifications to the Nydahl methodto make it applicable to the automated routine analysis of small samples ofnatural and polluted seawater, and to samples of contaminated saline effluents.

Nydahl [28] published a list of 39 compounds which gave oxidation yieldsover87%indistilledwater comparedwith theKjeldahlnitrogendetermination.He also included eight compounds which gave poor (< 83%) recoveries andconcluded that yields from peroxodisulfate oxidation are low for compoundscontaining nitrogen–nitrogen bonds and HN = C groups. The determinationof total dissolved nitrogen at the 10 µg N dm3 level for potassium nitrate, potas-sium nitrate, urea, glycine, 2, 2′-bipyridyl, and disodium EDTA added as spikesto natural seawater gave recoveries of 88–113% at a total dissolved nitrogenconcentration of 15.3 µg N dm–3. The detection limit was 0.18 µg N dm–3, i.e.23 ng N per sample.

Nakamura and Namiki [32] have determined total nitrogen in seawater byultraviolet spectrophotometry and correcting for background absorbance dueto bromide ion. Total nitrogen has been determined in seawater by oxidisingthe sample with potassium persulfate, boiling with sulfuric acid and potassiumpermanganate at pH 1.3 to remove bromide and bromate, removal of excesspermanganate and manganese dioxide with sodium thiosulfate and measuringthe absorbance of the treated solution at 220 nm [31].

Frankovitch and Jones [33] have presented a rapid automated method forthe determination of total combined nitrogen in seawater. The method hasa detection limit of 2 µM.

11.4Organic Phosphorus

The earlier methods for the measurement of organic phosphorus generallyused fairly drastic measures to decompose the organic compounds present andfree the phosphates. For example, Duursma [34] used the technique originallydescribed by Harvey [35] which included autoclaving the seawater with sulfuricacid at 140 ◦C for 6 h, at a pressure of 3 atmospheres. These methods required

11.5 Silicon 471

great care and a good deal of time; as a result, organic phosphorus was nevera routine determination.

Photo-oxidation was seen as a possible route to a total phosphorus method.Again, early work on the method was done by Armstrong et al. [15] andArmstrong and Tibbitts [36]. Grasshoff [37] adapted the method to continuousautomatic analysis; a variation on this method is considered the “standard”method for automatic analysis today [18]. Bikbulatov [38], on the other hand,feels that such important phosphorus compounds as ATP and DI are notcompletely decomposed by ultraviolet irradiation and that persulfate oxidationgives better results.

Both the persulfate and ultraviolet oxidation methods have much to recom-mend them. Ultraviolet photo-oxidation methods have an advantage in thatthey are easily automated.

The simultaneous determination of dissolved organic carbon and phos-phorus is feasible [39]. Phosphoglyceric acid (1 g equivalent P) was added tomembrane-filtered, pre-irradiated seawater, and the phosphate was measuredbefore and after ultraviolet irradiation using an autoanalyser. The recovery oforganic phosphorus was 100%.

Cembella et al. [40] have described a method for the determination of totalphosphorus in seawater. The procedure used magnesium nitrate to oxidiseorganic compounds before standard molybdate colorimetric determination ofortho-phosphate. The method was applied to several pure organic phosphoruscompounds and gave 93–100% recovery of phosphorus.

Ormaza-Gonzales et al. [41] compared methods for determining dissolvedand particulate phosphorus in seawater.

11.5Silicon

Silicon has been determined directly in seawater by inductively coupled plasmaatomic emission spectrometry with a detection limit of 0.3 µm silicon [42].

11.6Total Sulfur

Taylor and Zeitlin [43] described an X-ray fluorescence procedure for thedetermination of total sulfur in seawater. They studied the matrix effects ofsodium chloride, sodium tetraborate, and lithium chloride and show thatthe X-ray fluorescence of sulfur in seawater experiences an enhancement bychloride and a suppression by sodium that fortuitously almost cancel out.The use of soft scattered radiation as an internal standard is ineffective incompensating for matrix effects but does diminish the effects of instrumentvariations and sample inhomogeneity.


11.7Carbon Functions

The quantity of dissolved organic carbon in the oceans has been estimated tobe about 1018 g and constitutes one of the major reservoirs of organic carbon.Although large in total mass, the concentration of organic carbon in seawateris low (typically 0.5–1.5 mg C per litre).

The determination of organic carbon has always presented analytical diffi-culties. The content of inorganic carbon present in seawater is thirty or moretimes as great as that of organic carbon. To measure the organic carbon, eitherthe organic or the inorganic carbon must be removed from solution. The reten-tion of even a small percentage of the inorganic carbon could easily double ortriple the apparent organic carbon content of a sample. While this observationmight seem obvious, it is a fact that several workers have described methods inwhich the organic carbon was determined as the difference between a total andan inorganic carbon measurement, and where the measurement of inorganiccarbon had been incomplete. The resulting dissolved organic carbon valueswere too high by a factor of up to 10.

Even if removal of inorganic carbon is complete, the analysis of the remain-ing carbon is difficult. At a concentration of 1 ppm, a normal value for surfacewater in the open ocean, a 1 ml sample will contain 1 µg carbon. If we areinterested in differences between samples, we must strive for a precision of±5% or ±0.05 µg C per ml sample. This requirement places severe constraintson the sensitivity and precision of instrumentation.

Since any attempt to measure the productivity of a marine ecosystem musteventually require measurements of organic carbon in the various reservoirsof the system, an extensive literature exists on methods for the measurement oforganic carbon. The methods usually distinguish between total organic carbon(TOC), particulate organic carbon (POC), dissolved organic carbon (DOC),and volatile organic carbon (VOC). These are discussed below under separateheadings. A review of the available methods for all of these components up to1975 was published by Wangersky [44], and up to 1993 by Ogawa [45].

Both biochemical oxygen demand and chemical oxygen demand are mea-surements frequently made in water laboratories. Both of these methods aresubject to interference when applied to saline samples, as is discussed in theconcluding section in this chapter.

11.7.1Dissolved Organic Carbon

Because they are so intimately related, for the purposes of this discussion,the categories of “total” and “dissolved” organic carbon are combined in thissection. Also it is doubtful in any case that anyone ever measures a true“total” organic carbon in seawater. The total organic carbon should include

11.7 Carbon Functions 473

the dissolved, particulate, and volatile organic fractions; in the process ofremoving the inorganic carbon, usually by acidification and bubbling, someportion of the volatile organic carbon must be removed. Thus the total organiccarbon as measured by direct determination must always be the total minusthe volatile fraction.

The determination of dissolved organic carbon by oxidation methods inwater comprises three analytical steps: the removal of inorganic carbon fromthe sample, oxidation of the organic compounds to carbon dioxide, and thequantitative determination of the resulting carbon dioxide. The methods ofoxidation can be classified into three major groups:

1. Wet chemical oxidation methods, using oxidants such as persulfate, arewidely used in oceanographic and limnologic work [46,47]. The main draw-backs of these methods are their manual and cumbersome techniques andincomplete oxidation of some organic compounds [48].

2. High-temperature combustion, or dry oxidation methods developed forfresh and wastewater analysis, led to a whole range of discontinuous [49,50]and continuous [51, 52] commercial instruments. The discontinuous injec-tion method lack in sensitivity and show high blank values [53], whereasthe continuous combustion methods, with high sensitivity, do not toleratethe high ionic strength of seawater, even if designed for analysis of highlybuffered high-pressure liquid chromatography eluate [54].

3. The photo-oxidation of dissolved organic carbon in batch procedures [56,57] requires long reaction times, because of the thickness of the sam-ple. Automated, continuous procedures, where the sample is irradiated bya “medium pressure” mercury lamp [30, 57] brought great improvementand are utilised in several commercial instruments.

The merits and limitations of wet chemical oxidation, high-temperature com-bustion, and photo-oxidation methods for seawater analysis were summarisedby Gershey et al. [58].

The specifications for the future development of a sensitive, accurate, andautomated method were suggested by Wangersky [59].

The various methods used for the determination of total and dissolvedorganic carbon are now discussed under separate headings.

Ultraviolet Absorption

Armstrong and Boalch [60] have examined the ultraviolet absorption of sea-water, particularly in the wavelengths between 250 and 300 nm, where theabsorption is considered to result from the presence of aromatic compounds.Light absorption is a particularly useful measure, if it can be made to work,since it is not too difficult to construct an in situ colorimeter which can producecontinuous profiles of dissolved organic carbon with distance or depth [71].


Ogura and Hanya [62–65] investigated the components of the ultravioletabsorption in an attempt to devise a useful method for oceanic dissolvedorganic carbon measurements. They concluded that while the method mighthave limited application in coastal waters, most of the absorption in oceanicwaters was due to the inorganic components, principally nitrate and bromideions.

The method has since been revived for use in fresh water [65], where com-parisons with total organic carbon determinations by direct injection showedthe ultraviolet method to give results that were slightly high (3.8%), with a fairlylarge scatter (sd = 19% of the mean). Mattson et al. [66] applied a variant of themethod to coastal water, again calibrating against a direct injection total or-ganic carbon method, and claimed reasonable results. However, Wheeler [67]found that correlations between total organic carbon and ultraviolet absorp-tion, while quite strong over limited geographical regions, could switch frompositive to negative in adjacent regions.

Kulkarni [68] described an ultraviolet absorption method for the measure-ment of the organic carbon level in seawater which compares the intensities attwo wavelengths, one less than 350 nm and the other greater than 400 nm.

Biggs et al. [69] compared various instrumental methods for monitoringorganic pollution in water. These included total organic carbon, total oxygendemand, chemical oxygen demand, and biochemical oxygen demand, and anultraviolet method based on measurements of the ratio of the light transmittedby the sample at 254 nm in the ultraviolet to that at 510 nm in the visible region.The results obtained by the ultraviolet method were claimed to be largelyindependent of the presence of inert suspended solids in the sample. They alsoshowed that reasonably good correlations were obtained between ultravioletabsorbance at 254 nm on the one hand and total organic carbon, or biochemicaloxygen demand or chemical oxygen demand on the other. Although theseinvestigations were performed only on sewage effluents, this could be used asa rapid monitoring procedure for total organic carbon in seawater.

Wet Oxidation Methods

Another group of attractive techniques for determining total organic carbonare those using wet oxidation of the organic carbon to carbon dioxide. Inthese methods, the inorganic carbon is first removed by acidification and gaspurging to remove carbon dioxide produced by the decomposition of metalliccarbonates. The wet oxidation of organic carbon is then carried out by theaddition of the preferred oxidant, usually with heating, and the carbon dioxideresulting from the oxidation is measured by various methods.

Several different oxidants have been used in this work. The trend has beento stronger oxidants, in the hope that more complete oxidation would result.Among those used have been potassium peroxide [69], dichromate in sulfuricacid [70–73], silver-catalysed potassium dichromate [74, 75], potassium per-


sulfate [38], and silver-catalysed potassium persulfate [77,78]. Similarly, manymethods for measuring the evolved carbon dioxide have been devised, suchas titration of the oxidant consumed by coulometric titration (Duursma [74]),and conductometry and gas chromatography [72,76]. By far the most popular,however, has been the non-dispersive infrared gas analyser [78, 79].

Low-temperature ultraviolet-promoted chemical oxidation has providedmore reliable and precise data in the determination of low µg/l levels of totalorganic carbon [56, 78].

These instruments employ a continuous flow of persulfate solution to pro-mote oxidation prior to ultraviolet irradiation, and have a low system blankand low detection limit. Since all reactions take place in the liquid phase, prob-lems suffered by combustion techniques, such as catalyst poisoning, reactorcorrosion, and high-temperature element burnouts, are obviated. However,the ultraviolet-promoted chemical oxidation technique is not designed to han-dle particulate-containing samples, and tends to give incomplete oxidation forcertain types of compounds such as cyanuric acid.

While the precision of these methods, of the order of ±0.1 mg C per litre, isnot high enough for purposes of maintaining budgets on various fractions ofthe total organic carbon, it is certainly good enough to provide some idea ofworld-wide distributions. The great disadvantage of all wet oxidation methodsis theuncertaintyof completenessofoxidation.Eachchange toamorepowerfuloxidanthasbeenmadewith the implicit assumption that thisparticularoxidantwould finally work on all of the compounds and give total organic carbon valuesequivalent to total combustion values. In the absence of a recognised refereemethod, the only checks on the completeness of oxidation have come from theoxidation of known pure compounds and from the oxidation of 14C-labelledmaterial derived fromplanktoncultures.Bothof these techniques seemto showessentially total oxidation by the high-temperature persulfate method [47].

There is considerable controversy concerning the completeness of the wetoxidation. Some of this controversy has centred on the calculation of the blankvalue for the method. One method for the preparation of organic-free waterconsists of the oxidation of the residual organic matter with phosphoric acidand persulfate in the process of distillation. This procedure should produce wa-ter that is immune to further oxidation with the same reagents, and thereforeproduce zero blanks. However, many workers using total combustion tech-niques have found that small but measurable amounts of organic carbon areresistant to this, or to any other, wet chemical oxidation. This is not surprising,since there are a number of organic compounds that are not completely oxi-dised even under much more stringent conditions, such as boiling with nitricand perchloric acids for periods of up to 45 min and at temperatures up to203 ◦C [80].

While there have been very few intercalibration studies done between thevarious kinds of methods, the few available comparisons suggest that boththe ultraviolet irradiation method and the various total combustion methods


discussed in later sections find more organic carbon than do the chemical wetoxidation methods.

However, the work of Menzel and Vaccaro [46] and others [15, 81] suggeststhat wet oxidation is useful for determining the dissolved organic content ofseawater samples, and the oxidation has been shown to be essentially com-plete [82]. Seawater is first freed of inorganic carbon by treatment with a smallvolumeof3%phosphoric acid, and theorganic carbon is thenoxidised in sealedglass ampoules in an autoclave at 130 ◦C using potassium persulfate as an oxi-dant. The resulting carbon dioxide is passed through a non-dispersive infraredanalyser whose signals are related to milligrams of carbon in the sample.

Hannaker and Buchanan [82] used a method based on wet oxidation withpotassium persulfate [83] for the determination of dissolved organic content inconcentrated brines following the removal of inorganic carbonates with phos-phoric acid. The method involves wet oxidation with potassium persulfate at130 ◦C followed by a hot copper oxidation and gravimetric measurement of thecarbon dioxide produced. The technique overcomes difficulties of calibrationcurvature, catalytic clogging, and instrument fouling often encountered withinstrumental methods.

Low-volatility natural organic material such as polysaccharides and highermolecular weight proteins sometimes produced low results. In the Hannakerand Buchanan method [82] these problems are overcome by using a solution-phase oxidant and enclosing the system in a sealed tube. In this way all ofthe constituents are fully contained and exposed to oxidation and, moreover,oxidation of the organic matter to carbon dioxide is complete for the greatermajority of compounds.

Hannaker and Buchanan [82] differ from Menzel and Vaccaro [46] in thatthey use Carbosorb or soda asbestos tubes to estimate the carbon dioxideproduced, instead of the non-dispersive infrared analyser used by the latterworkers.

Using this method, a series of organic species were measured at variousconcentrations.

The results presented in Table 11.1 show that very close agreement wasobtained for a variety of organic compounds both with and without sodiumchloride present.

Ultraviolet Irradiation Methods

Another promising wet oxidation method uses high-intensity ultraviolet lightin the presence of an oxidising agent such as hydrogen peroxide or potassiumpersulfate. This method was proposed by Beattie et al. [55], who used a massspectrometer as a detector for the carbon dioxide produced. Armstrong andhis co-workers adapted the method to seawater and showed that the methodgave essentially the same results as the persulfate oxidation [15,23]. They notedthe incomplete oxidation of urea, using this technique.


Table 11.1. Analysis of a series of organic species at various concentration levels, with andwithout NaCl present [82]

Organic compound Range tested, NaCl range, No. of samples Percentagetested ppm ppm tested mean recovery

Mannitol 54–432 0–300 10 98.4Acetic acid 30–720 0–300 11 97.9D-tartaric acid 76–604 0–300 9 102.7Oxalic acid 57–1140 0–300 7 100.9Malic acid 36–716 0–300 9 99.0Propan-2-ol 45–180 0–300 4 94.4Acetylacetone 232–465 0–300 4 99.8Ethanol 208–417 0–300 4 98.1Glycine 10–382 0–300 10 95.5Glycerol 10–392 0–300 9 98.7Trisodium citrate 123–490 0–300 7 96.0Salicylic acid 42–406 0–300 9 98.3

Williams [81,84] found that ahigh-energyultraviolet oxidationmethodgavedissolved organic carbon results for seawater samples that were higher thanthose obtained by the wet persulfate oxidation method described by Menzeland Vaccaro [46]. Other variations on the method have been published.

One of the great advantages of this method is that it is easily adapted forautomatic analysis [30,57,78]. Ehrhardt [57]detectedgeneratedcarbondioxideby a conductometric method, Collins and Williams [30] used an infraredanalyzer, and Armstrong and Tibbits [23] and Goulden and Brooksbank [78]used a continuous photoelectric method. A disadvantage of these automaticmethods is that owing to long irradiation times, there is a considerable lagbetween the intake of the first sample and obtaining the first result.

An example of an automated colorimetric method, oxidation by ultravioletirradiation for the determination of dissolved organic carbon in the presence ofpotassiumpersulfate [87], is thatof Schreurs [86].Themethoduses aTechniconanalyser to measure dissolved organic carbon, is fast and precise, and may beused to measure dissolved organic carbon in both seawater and fresh waterover the range 0.1–10 mg C per litre.

In this method the sample is acidified and the inorganic carbon is removedwith nitrogen. An aliquot is resampled for analyses. Buffered persulfate isadded and the sample is irradiated in the ultraviolet destructor for about 9 min.The hydroxylamine is added and the sample stream passes into the dialysissystem. The carbon dioxide generated diffuses through the gas-permeablesilicon membrane. A weakly buffered phenolphthalein indicator solution isused as the recipent stream, and the colour intensity of this solution decreasesproportionately to the change in pH caused by the absorbed carbon dioxide


gas. The measurement of the colour intensity is done in a 15 mm flow cell at550 nm.

Linear response to within 0.1% of the theoretical least-squares fit line isobtained with the intercept passing through zero.

Recoveries were obtained for a range of organic compounds represent-ing various organic classes and containing a variety of functional groups.These organic compounds with a concentration of 2 and 4 mg C per litre weredissolved in standard seawater. Recovery was measured against potassiumhydrogen phthalate standards. The average recovery is 98.9%. Mueller andBandaranayake [39] reported on an automated method for the determinationof dissolved organic carbon in seawater using continuous thin-film ultravioletoxidation in a quartz spiral ultraviolet “low-pressure” mercury lamp. Testswith a range of organic compounds showed that the instrument is capable ofrapid, continuous, and reliable dissolved organic carbon analysis in seawater,and is suitable for use aboard a research vessel.

Mueller and Bandaranayake [39] were able to show that more than 95% ofthe following compounds were oxidised in the first run, when present in thewater sample at the 5 mg C per litre level: oxalic acid, potassium phthalate,humic acid, glucose, sucrose, ascorbic acid, glycine, and phenol. Only sulfurcompounds gave incomplete recoveries [58, 88].

A pH of 3 was optimal for the complete removal of inorganic carbon andthe most efficient oxidation of nitrogen-free organic compounds, while a pHof 2.5 was optimal for nitrogenous compounds (Fig. 11.1).

These workers found that the efficiency of oxidation was a function of theresidence time of the sample in the reactor and the flow rate of the carrier gas.A high precision of carbon dioxide determination was achieved at a sampleflow rate of 50 ml/h and a carrier gas flow rate of 62 l/h, of which 40 l/h passesthrough the shielded zone.

Using this procedure, analysis can be completed in 5–10 minutes.

Evaporation–Dry Combustion

It seems that in principle the simplest of all possible total organic carbonmethods would use acidification of the sample followed by evaporation anddry high-temperature combustion of the resulting sea salts. This method hasbeendevelopedboth for freshwater [89] andseawater [90].Thedata foroceanictotal organic carbon resulting from the latter method were considerably higher,by a factor of 2 or 3, than those coming from the wet oxidation methods. Thisdifference was later ascribed to contamination of the samples by the apparatusand impurities in room air.

A method using vaporisation at low temperature in an atmosphere oforganic-freegas, anextensionof themethodofSkopintsevandTimofeyeva [90],was developed by MacKinnon [91]. Contamination from organic vapours inthe laboratory air was a major problem; as soon as this was understood, and


Figure 11.1. Effect of pH on the oxidation of organic compounds. 1 Oxalic acid. 2 Potassiumphthalate. 3 Ascorbic acid. 4 Humic acid. 5 Phenol. 6 Guanidine. 7 Glycine. 8 Urea. 9 Glucose.10 Sucrose. Nitrogen-free compounds peak at pH 3, nitrogen-containing compounds atpH 2.5. The seawater sample was collected from Myrmidon Reef, 80 km north of Townsville,Australia. Source: [39]

sufficient care was given to safeguarding the evaporation process, both accu-racy and precision reached reasonable levels. Comparisons were run betweenthis dry combustion method and the Sharp [58] version of the persulfate wetoxidation method. The dry combustion method gave results consistently 15–25% higher than persulfate wet oxidation. On water from the same parts ofthe ocean, the values were well below those found by Skopintsev and Timo-feyeva [90] and Gordon and Sutcliffe [29].


MacKinnon [92] has described a high-temperature oxidation method forthe accurate and precise determination of the total organic carbon in seawa-ter. Problems of contamination in sample storage, preparation, and oxidationwhich are evident in previous dry oxidation methods were controlled. Thetotal organic carbon results from different areas are determined and com-pared directly with the results obtained by the persulfate oxidation method.A high correlation between the two methods was obtained. In this method, theorganic matter is oxidised to carbon dioxide, in an oxidation similar to thatof Skopintsev [93], after a sample preparation similar to that of Gordon andSutcliffe [29]. The oxidation products are determined with a non-dispersive in-frareddetector.Valuesobtainedby thisprocedureare similar to the lowervaluesobtained by Sharp [58] with his high-temperature combustion method. In themethod by Gordon and Sutcliffe [29] the aqueous sample is evaporated to dry-ness by freeze-drying, and the resulting salt is oxidised in a high-temperaturefurnace.

MacKinnon [92] carried out a very detailed study in which sample col-lection and storage, sample preparation, the dry oxidation procedure, watercorrections, adsorption effects, and precision and accuracy are all discussedin detail.

Comparison of Results Obtained by Evaporation–Dry Combustionand Ultraviolet Photo-oxidation

Initially, results reported for dry-combustion methods were found to be higherthan wet-oxidation methods based on persulfate by factors of 2 or more. Thisdiscrepancy has steadily decreased as methodologies have improved. Contam-ination problems of dry methods have been reduced and the oxidation effi-ciency of the wet methods has been improved. While the differences betweenapproaches have been discussed [94, 95], there is still uncertainty whether theremaining difference between the two techniques is real – a result of incompleteoxidation, incorrect estimation of blanks, or a combination of both.

Collins and Williams [30] have recently described a modification of Ehr-hardt’s earlier photochemical method [57], which offers the practical advan-tages of speed, convenience, and the potential for “real-time” analyses. How-ever, until the accuracy of the results is established, the method will not receivegeneral acceptance. Collins and Williams [30] examined the completeness ofoxidation of their photo-oxidation system using three independent methods,but pointed out that while essentially complete oxidation was indicated, defini-tive proof was lacking. A more satisfactory solution to the problem might befound through comparison of results of the photo-oxidation method with thedry-combustion method, which most analysts are willing to accept as com-plete [96].

In this connection, Gershey et al. [58], in a detailed study, have compared re-sults obtained for dissolved organic carbon in seawater using the evaporation–


dry combustion method of MacKinnon [92], the ultraviolet photochemicalmethod of Collins and Williams [30] (slightly modified), and the persulfateoxidation method of Menzel and Vaccaro [46] and Sharp [48].

Gershey et al. [58] have shown that the agreement among the three tech-niques for analysis of dissolved organic carbon in seawater is generally good,considering the problems inherent in this type of trace analysis. Taking their re-sults and thoseofGouldenandBrooksbank [78] together, one canconclude thatthe difference between the results of the dry combustion and photo-oxidationmethods is small and of little practical consequence.

The versions of the persulfate oxidation methods used at present yieldresults that are lower than those obtained using the dry combustion or photo-oxidation techniques (Table 11.2). Close agreement between the persulfate andother methods is obtained when the analyses are carried out on freshwaterrather than seawater samples. If the persulfate oxidation procedure is to beretained as a method for seawater analysis, serious consideration should begiven toabandoning thepresentbatchwiseprocedure in favourof anautomatedprocedure.

Table 11.2. Comparison of the different oxidation methods for analysis of organic carbonin filtered and unfiltered seawater

Area n Averaged organic carbon concentration(mg C per l)

Difference bypaired t-test at95% confidencelevel

Dryoxidation[105]

Photo-oxidation[61]

Persulfateoxidation[49]

Scotian shelf 25 – 0.85 ± 0.14 0.77 ± 0.13 S(0–500 m) 26 0.88 ± 0.13 0.85 ± 0.14 – NS(31–35‰) 25 0.88 ± 0.13 – 0.77 ± 0.13 S

Coastal area

(a) filtered(DOC)

9 – 1.26 ± 0.06 1.12 ± 0.10 S

9 1.31 ± 0.08 1.26 ± 0.06 – NS9 1.31 ± 0.08 – 1.12 ± 0.10 S

(b) unfiltered(TOC)

5 – 1.41 ± 0.15 1.38 ± 0.21 NS

5 1.68 ± 0.19 1.41 ± 0.15 – S5 1.68 ± 0.19 – 1.38 ± 0.21 S

S: SignificantNS: Not significantDOC: Dissolved organic carbonTOC: Total organic carbonSource: [58]


Gershey et al. [58] concluded that a continuous and automated photo-oxidation procedure of the type described by Collins and Williams [30], withthe reported modifications, will probably satisfy most of the needs of theoceanographer concerning measurement of dissolved organic carbon. Theconvenience and rapidity of the method opens up a new area of research: thestudy of the small-scale temporal and spatial variations of the dissolved organiccarbon content of the oceans.

Dry Combustion Direct Injection Methods

These procedures differ from the evaporation–dry combustion proceduresdescribed in the previous section as follows.

Evaporation–Dry Combustion

Inorganic carbonate is removed by acid addition, evaporated at fairly lowtemperature to a dry sea salt residue, and the residue combusted to carbondioxide in a high-temperature furnace.

Dry Combustion–Direct Injection

Inorganic carbonate is removed by acid addition, and the acidified liquidsample is injected directly into a high-temperature furnace.

The dry combustion–direct injection technique provides many advantagesover other methods, such as quick response and complete oxidation for de-termining the carbon content of water. Its primary shortcoming is the needfor rapid discrete sample injection into a high-temperature combustion tube.When an aqueous sample is injected into the furnace, it is instantaneouslyvapourised at 900 ◦C and a 5000-fold volume increase can be expected. Sucha sudden change in volume causes so-called system blank and limits the max-imum volume of injectable water sample, which in turn limits the sensitiv-ity [106, 107].

A dry combustion-direct injection apparatus was applied to water samplesby Van Hall et al. [51]. The carbon dioxide was measured with a non-dispersiveinfrared gas analyser. Later developments included a total carbon analyser [97],a diffusion unit for the elimination of carbonates [98], and finally a dualtube which measured total carbon by combustion through one pathway andcarbonate carbon through another. Total organic carbon was then calculatedas the difference between the two measurements [99].

Van Hall et al. [100] inject a 20 litre sample into a high-temperature furnaceat 950 ◦C containing catalyst to promote oxidation of carbon compounds tocarbon dioxide, which is then passed into a non-dispersive infrared analyser.The carbonate interference can be determined by passing an acidified portionof the sample through a low-temperature furnace [101–103].


One of the best-known commercial instruments developed for organic car-bon determinations is the Beckman total carbon analyser, which utilises ananalysis scheme developed by Van Hall and co-workers [57, 99]. This instru-ment works reasonably well in fresh water. It has become a standard instrumentin pollution control and water treatment [103]. The Beckman instrument hasnot worked as satisfactorily for seawater because of the latter’s high carbonateand low organic content.

Another instrument developed by the Precision Scientific Co. was basedupon the work of Stenger and Van Hall [99, 104]. Experience has shown thatapplication of the Beckman and Precision Scientific Co. instruments to con-centrated or saturated brine solutions leads to erratic and unreliable results.There are several possible reasons for this: (1) the catalyst will rapidly becomecoated with sodium chloride; (2) oxidation of Cl– to chlorine will occur; and(3) volatile organics may not all be trapped by the solid catalyst.

Van Hall, Safranko, and Stenger [51] have also pointed out that strong brinesinterfere with the method by producing “fogs” which may be counted as carbondioxide, while in cases where the flame ionisation detector is being used, largespikes appear in the recorded curve [105].

Sharp [48] has described a dry combustion-direct injection system built foroceanographic analyses. This unit used 100 µl samples, injected into a 900 ◦Coven in an atmosphere of oxygen. The output from a non-dispersive infraredcarbon dioxide analyser was linearised and integrated.

Ton and Takahashi [108] describe a dry combustion-direct injection to-tal organic carbon analyser, the DC-90 (produced by Dohrmann Division,Xertex Corporation), which is claimed to overcome many of the problemsencountered with earlier versions of this equipment. In this analyser, car-rier water is continuously pumped into a ceramic combustion tube wherea constant flow of oxygen is provided. The combustion tube is designed sothat the carrier water is dispersed and instantaneously evaporated. A sam-ple is introduced by either a rotary injection valve (via a sample loop) or bysyringe into the carrier water stream. Thus gas and stream are maintainedat constant flow. Carbonaceous materials in the sample are completely oxi-dised to carbon dioxide in the presence of oxygen, an oxidation catalyst, anda heat transfer device. After the stream is condensed, carbon dioxide in oxy-gen is subsequently measured by non-dispersive infrared spectroscopy. Thecarbon dioxide generated by complete combustion is directly proportionalto the total carbon in the stream. Since a large sample volume can be in-jected without disturbing flow conditions, the system blank is negligible; thushigh precision and sensitivity can be achieved. This continuous water carrierflow–high-temperature combustion technique works quite well for determin-ing total organic carbon in samples such as seawater and brine solutions. Theconstant water flow prevents salt build-up, and deterioration of the combus-tion tube is eliminated by the use of a ceramic tube. Inorganic carbon canalso be directly determined by injecting a sample into a gas/liquid separa-


tor. Total organic carbon can be determined either by externally acidifyingand sparging the sample before analysis, or it can be calculated by taking thedifference:

Total organic carbon = total carbon – inorganic carbon

Shimadzu supply a fully automated total organic carbon analyser, theirModel TOC 500, based on dry combustion-direct injection followed by non-dispersive infrared spectroscopy. The system incorporates a microcomputerand has an optional automatic sample injector. The method is applicable toall waters, including seawater, and has a range of 1–3000 ppm. It is also ap-plicable to the determination of volatile organic carbon. In this instrumentflow-controlled and humidified carrier gas (highly pure air) is allowed to flowthrough the TTCU (total carbon) combustion tube, which is kept at 680 ◦C.Then the gas flows through the drain separator, where its moisture is removed,through the inorganic carbon reaction tube kept at 150 ◦C, and through theelectronic dehumidifier, where its moisture is removed again. Finally, the gasflows through the cell of the non-dispersive infrared analyser, which measuresthe carbon dioxide concentration in the gas.

When a sample is injected into the total carbon injection port with a mi-crolitre syringe, the carbon atoms in the organic and inorganic compoundsin the sample are oxidised into carbon dioxide. The NDIR analyser generatesa peaked signal with an area proportional to the carbon dioxide concentra-tion. The built-in microcomputer measures the peak area and converts it intoa total carbon concentration value by means of the calibrating equation pre-determined through measurement of standard solution samples.

When a sample is injected into the inorganic carbon injection port, the sam-ple enters the inorganic carbon reaction tube, where only the inorganic carbonis turned into carbon dioxide. The carbon dioxide concentration is convertedinto inorganic carbon concentration value in the same way as total carbon.The total organic carbon concentration is obtained by subtracting inorganiccarbon concentration from total carbon concentration and printed out. Withthe optional volatile organic carbon measuring channel added, volatile organiccarbon concentration is measured in the same way.

When the volatile organic carbon measuring channel is selected, the gascoming out of the total carbon combustion tube enters the volatile organiccarbon evaporation tube, which is packed with volatile organic compoundevaporation reagent and inorganic carbon absorbing reagent, and is kept at150 ◦C. Then the gas flows through the heated tube into the volatile organiccarboncombustion tubekeptat680 ◦C,andreaches thenon-dispersive infraredanalyser via the electronic dehumidifier. When a sample is injected into thevolatile organic carbon injection port, only the volatile organic componentsare vapourised. Some inorganic carbon components generate carbon dioxide,which is absorbed by the inorganic carbon absorbing reagent. Volatile organic


components are oxidised in the volatile organic carbon combustion tube intocarbon dioxide. The carbon dioxide concentration is converted in the sameway as inorganic carbon and printed out.

Dry combustion-direct injection techniques using gas chromatography asthe method of measurement have been devised for fresh water and low salinitywater, but not for seawater. Nelson and Lysyj [109] used pyrolysis of the organiccompounds without oxidation, with flame ionisation as the detection method.The technique was further developed by Eggertsen and Stross [105]. Since theresponse of this detector varies according to the type of organic compoundpresent, the accuracy of the method cannot match its precision. This prob-lem was overcome [110–113] by injecting the sample into a high-temperaturefurnace containing cupric oxide so that the organic components are oxidisedto carbon dioxide, then reducing the carbon dioxide to methane, which ismeasured with a flame ionisation detector.

A more recent development is the Dohrmann DC-54 Ultra Low Level to-tal organic carbon analysis. This equipment is capable of determining totalorganic carbon down to 10 µg/l, and purgeable organics down to 1 µg/l. It isapplicable to seawater. The principle employed is ultraviolet-promoted chem-ical oxidation of organic carbon to carbon dioxide, followed by conversion tomethane, which is determined by flame ionisation gas chromatography. Anal-ysis time is 8–9 min, and the range of application 0–10 000 µg/l. The precisionfor total organic carbon is ±1 µg/l.

Preservation and Storage of Samples for the Determination of Dissolved OrganicCarbon

As was expressed by Wangersky and Zika [96], sampling and storage of watersamples for the determination of organic compounds can be an importantsource of errors. Duursma [34] stored his samples in 500 ml glass bottles,preserving with sulfuric acid. Wangersky and Zika [96] claim that addition ofacid to samples can change organic compounds.

Merks and Vlasbom [114] filtered 20 ml sample through a glass fibre filterand then transferred the solution into a 20 ml glass ampoule. Immediatelyafter that the ampoule was closed by heating it in a flame. The filters werepre-treated by heating at 330 ◦C, and the ampoules by heating at 610 ◦C, bothfor 2 h. After closing, the ampoules were deep frozen at –20 ◦C. Just beforethe determination of dissolved organic carbon the ampoule was thawed againand opened. Results of several compared analyses carried out by Olrichs [115]showed very good agreement using different analytical systems for measuringtotal organic carbon. However, Elgershuizen [116] did not find good agreementbetween determinations of dissolved organic carbon which were carried outin two laboratories. As the sampling took place at the same time and in thesame way, the cause of the bad results can only be a matter of preservation andstorage.


In fact, the conditions of storage were very different. Merks and Vlas-bom [114] used glass ampoules stored at –20 ◦C, whereas Elgershuizen [116]stored the samples in PVC bottles at 4 ◦C, acidified to pH 2.

Merks and Vlasbom [114] carried out some comparative experiments onstandard seawater with potassium hydrogen phthalate and water from EasternScheidt and Western Scheidt. These three types of samples were stored in threedifferent ways and in three types of storage bottles.

The storage occurred in a deep-freezer at –20 ◦C, in a refrigerator at 4 ◦Cand at room temperature at about 20 ◦C in the laboratory. The bottles used inthis investigation were glass ampoules, hard plastic bottles, and polystyrenesample cups as used in autoanalysers. The pretreatment of those bottles wasas follows: the glass ampoules were heated at 610 ◦C for 2 h, the plastic bottleswere flushed several times with deionised water from a Millipore Milli-Q outfitand dried afterwards. The polystyrene cups were not pretreated at all. Allsamples were filtered over preheated glass fibre filters just before the threetypes of bottles were filled. Before filling with the sample, the plastic bottlesand the polystyrene cups were rinsed twice with the filtered sample, while theglass ampoules were filled without rinsing.

From the results it can be concluded that the decrease in dissolved organiccarbon with storage time is always lowest when the samples come deep frozen;deep freezing in a glass ampoule resulted in the lowest decline. Even underthese conditions the storage time should be as low as possible.

Miscellaneous Methods for Measurement of Dissolved Organic Carbon

Methods that have beenapplied only to fresh water samples show somepromisefor seawater analysis. One such system [117] measures inorganic, volatile or-ganic, and nonvolatile organic carbon fractions on the same sample. Thedifferent fractions are evolved separately, the inorganic carbon by acidifica-tion with phosphoric acid, the volatile fraction by purging with oxygen andhigh-temperature combustion of the vapours, and the nonvolatile fraction byultraviolet photo-oxidation. In each case, the carbon dioxide evolved is purifiedby cold trapping.

Another technique depends upon the difference in rates of volatilisation todifferentiatebetweenorganic and inorganic carbon, anduponplasmaemissionspectrography for the final measurement [118]. The sample is dispensed intoa platinum boat, dried, and then covered with an oxidant, in this case V2O5.The sample is then moved into an oven held at 850 ◦C. At this temperaturethe organic carbon is volatilised a few seconds before the carbonate carbon.Light from a microwave-excited argon plasma, into which the volatile materialis passed, is dispersed by a monochromator, and the 193 nm atomic carbonline is measured with a photomultiplier. An advantage of this technique is thatoxidation to carbon dioxide need not be complete. When comparisons wererun between this method and persulfate oxidation, no significant differences


were obtained. The obvious difficulty to be expected in using this technique forseawater analysis stems from the vastly greater concentration of carbonates.The peaks for organic and inorganic carbon are not completely separated. Evena small overlap between the peaks would prove disastrous.

Mills et al. [119] carried out reversed-phase liquid chromatographic stud-ies of dissolved organic matter and copper-organic complexes isolated fromestuarine waters.

The adsorption of organic matter on any surface presented to seawaterhas been well documented. Neihof and Loeb [120] have demonstrated thisadsorbance by following the change in surface charge of newly immersedsurfaces. There has even been an attempt to use this phenomenon as a meansof measuring dissolved organic carbon. Chave [121] found an associationbetween calcite and dissolved organic materials in seawater, and Meyers andQuinn [122] tried to use the effect as a method for the collection of fatty acids.As a collection technique, adsorption on calcite has several advantages. ThepH of the sample is not greatly altered by the addition of small amounts ofcalcite; the precipitate is dense and should settle quickly; and after filtrationthe inorganic support can be removed by acidification. Unfortunately, therecovery of added fatty acids was inefficient – of the order of 18%. Meyers andQuinn [123] achieved a somewhat greater efficiency of collection of fatty acidswith clays, but the insolubility of the clays nullified one of the advantages ofthis concentration technique.

Ferric hydroxide is the precipitant most commonly used for the collectionof organics. It is formed by the in situ formation of hydrated ferric oxides,usually by the addition of ferrous iron, followed by potassium hydroxide. Thetechnique was first used for the precipitation of organic matter from an agedalgal culture [124]. They recovered 79–95% of the 14C-labelled material fromsuch cultures. Williams and Zirino [125], measuring efficiencies of removalof dissolved organic carbon, found that such scavenging collected between 38and 43% of the organic carbon measurable by wet oxidation with persulfate.Chapman and Rae [126] examined the effect of this precipitation on specificcompounds. They found coprecipitation to be more complete with copper hy-droxides, but still far from satisfactory. Only certain compounds were removedeffectively by this treatment, and the efficiency of removal varied with the watertype and the organic compound involved.

11.7.2Dissolved Inorganic Carbon

The strong increase in atmospheric concentrations of carbon dioxide [127]has generated considerable interest in the global carbon cycle [128–130]. Tech-niques for determining the components of the carbonate system have beenrefined, new techniques have been developed, or both. Among the four mea-surable parameters (total inorganic carbon), pH, pCO2, and total alkalinity


of the carbonate system, the total inorganic carbon has a typical range of1900–2450 µM. TIC is routinely measured with high precision on semiauto-mated equipment by coulometry [131, 132], which is the generally acceptedmethod for oceanographic research. Calibration is performed by standardprocedures [133, 134] and certified reference materials as supplied by Dick-son [134]. Other methods are determination via infrared detection followingSugimura and Suzuki [135] and Wiebinga [136], and by conductivity [137].The former determines both total organic carbon by catalytic oxidation andtotal inorganic carbon by injection onto a low-temperature column, whichdoes not oxidise any organic matter present. The latter does the same, albeittotal inorganic carbon is determined via acidification of the sample. Afterdiffusion across a silicone membrane in a dialyser, the increase in conduc-tivity is determined, which is proportional to the amount of total inorganiccarbon.

Biogeochemical processes, and especially their rates, modify the char-acteristics of the water column [138]. The introduction of underway mea-surements of carbonate system parameters in combination with tempera-ture/salinity/fluorescence provided insight into the hitherto unexpected highvariability and inherent deduced rates found in the surface ocean [139, 140].These data also laid the foundation for a now quite extensive database usedfor modeling purposes as well as “ground truth” data for satellite obser-vations (GOOS, Global Ocean Observing System; TOGA-COARE, TropicalOcean Global Atmosphere-Coupled Ocean-Atmosphere Response Experi-ment). Though temperature and salinity gradients in the surface waters areresolved on a routine basis, resolving the spatial scale of chemical tracers insurface waters is typically limited either by the sample acquisition process(e.g., a station) or by the rate at which samples can be processed on boardship. Recent advancements have been made in shorter analysis times, butfor total inorganic carbon via coulometry this has been only modest, as theelectrochemical titration cannot be faster without losing its accuracy.

Hall and Aller [141] described another method based on flow injectionanalysis for determining total inorganic carbon and NH+

4 in both marinewaters and freshwaters. Standards for the total inorganic carbon analysis wereprepared using sodium bicarbonate in distilled water. For a range of 0–2 mM(the latter being the upper limit in seawater), a RSD of 1.4% was found, whichequals ∼28 µM. A major advantage of their method was the speed at whichsamples could be analysed. A drawback, from an oceanographic point of view,is that the precision was less than required to resolve the fine details in theoceanic water column as needed for carbon budget.

Stoll et al. [142] have described a rapid continuous-flow determination oftotal inorganic carbon in seawater samples. The method runs on an autoanal-yser Traacs 800 spectrophotometric system and is calibrated versus certifiedreference materials readily available. A typical analysis speed of 45 samples perhour can be reached with an accuracy of 2–3 µM and a precision of ∼2.5 µM.


The analysis requires only a small amount of sample and is thus ideally suitedfor pore water samples and samples taken from cultures where sample volumeis at a premium. The speed of the analysis makes mapping of oceanic surfacewater characteristics possible. Potential interference of sulfide in anoxic (e.g.pore water) samples can be masked by the addition of a hydrogen peroxidestep. Although the latter is a strong oxidative reagent, no significant effecton total inorganic carbon concentration due to oxidation of (labile) organicmatter could be found.

11.7.3Particulate Organic Carbon

The simplest of the components of the organic carbon system to measure is theparticulate phase, since the solid material can be isolated by filtration. Once thesampling conditions for filtration have been defined, the analytical proceduresare straightforward. The major problem is that of sensitivity; depending uponthe filter used, the particulate organic carbon content of surface seawater willrun between 25 and 200 µg C per litre, while the deep water will give values of3–15 µg C per litre. Almost without exception, the modern methods for par-ticulate organic carbon use a high-temperature combustion of the filter andits organic load. Many early workers employed a wet oxidation with chromicand sulfuric acids, with the actual measurement being a titration with ferrousammonium sulfate [143]. A rough conversion from wet oxidation to dry com-bustion values can be achieved by multiplication of the wet oxidation valuesby a factor of 1.09, commonly used for this purpose.

The various combustion methods differ primarily in the method of mea-suring the carbon dioxide generated from the organic carbon. The first reallysensitive carbon dioxide detector and the one still most used is the non-dispersive infrared gas analyser. The detecting element senses the difference inabsorption of infrared energy between a standard cell filled with a gas with noabsorption in the infrared, and a sample cell. Water vapour is the only seriousinterference, hence the carbon dioxide must be dried before any measurementsare made.

The carbon dioxide resulting from combustion of the particulate organiccarbon can be passed through the analyser, producing a single spike ona recorder, or it can be pumped in a loop through the analyser, until an equi-librium concentration is reached in the loop. Since the output of the analyseris nonlinear, the latter technique has been favoured by some investigators [46].

Carbon analysers using the non-dispersive infrared analysers have beendescribed by Kuck et al. [144] and by Ernst [145], among others.

Electrometric methods have also been used for the final measurement ofcarbon dioxide; Szekielda and Krey [146] devised a conductometric method.By far the most popular methods, however, have used commercial CHN analy-sers for the final measurement [147–150]. The actual measurement of carbon


dioxide with these instruments is made by gas chromatography; Perkin-Elmersupply a CHN analyser.

Of course, since these analysers measure all of the carbon dioxide generated,any carbonate present in the sample will also be measured as organic carbon.Many investigators remove any carbonate carbon by fuming the filters withhydrogen chloride vapour, or by treating them with dilute hydrochloric acid.Differential combustionhas also been suggested as a method for distinguishingbetween carbonate carbon and organic carbon; baking at 500 ◦C for 4 h wasused to remove organic carbon from carbonate-rich particulate matter [151].

An alternative method for the determination of particulate organic carbonin marine sediments is based on oxidation with potassium persulfate followedby measurement of carbon dioxide by a Carlo Erba non-dispersive infraredanalyser [152, 153]. This procedure has been applied to estuarine and high-carbonate oceanic sediments, and results compared with those obtained bya high-temperature combustion method.

Weliky et al. [154] described a procedure for the determination of bothorganic and inorganic carbon in a single sample of a marine deposit. Carbonatecarbon is determined from the carbon dioxide evolved by treatment of thesample with phosphoric acid; the residue is then treated with a concentratedsolution of dichromate and sulfuric acid to release carbon dioxide from theorganic matter. The carbon dioxide produced at the two stages of the analysis isestimated using a carbon analyser based on the thermal conductivity principle.In addition, total carbon content is determined on another subsample usingthe dry combustion furnace. This provides a check on the values determinedby the phosphoric acid dichromate technique.

Salonen [155] compared different glass-fibre and silver-metal filters for thedetermination of particulate organic carbon. He puts forward a theory that thevarying characteristics of different filters, or pore sizes, appreciably modify theresults of particulate organic carbon determinations, making comparison ofpublished concentrations unreliable. Salonen compares silver-metal filters anddifferent types of glass-fibre filters, and seeks to find a filter which would havebiologically meaningful cut-off size of particles. The most retentive glass-fibrefilters were able to retain almost all bacteria from the water of an oligotrophiclake; they prove to be quite near to the ideal. Silver filters provide similarretention, but because of their high blank values, price, and lower filtrationspeed and capacity, they are not able to compete with glass-fibre filters inpractical work.

11.7.4Dissolved Organic Carbon

Many volatile organic compounds (hydrocarbons, alcohols, aldehydes, acids,esters, ketones, amines, etc.) have been identified in marine systems [156,157].These volatile materials may have an important role in the cycling of organic


matter in natural waters. Volatile or low molecular organic materials may beproduced in situ by biological [158–161] and chemical reactions [162–164] orcan be introduced into the marine systems by human activity [165, 166], orthrough fluvial and atmospheric transport.

Direct methods of analysis such as distillation [158,167,168], liquid–liquidextraction [159,169], headspace analysis [170–172], dynamic headspace analy-sis [157,173–178], and direct injection [179] have been used mainly for specificvolatile components.

As for all of the fractions of organic material in seawater, the volatile organiccarbon fraction is defined by the method by which it is collected. In one of theearliest estimates, Skopintsev [93] defined the volatile fraction as the differencebetween total organic carbon values, as measured by evaporation and drycombustion, when the evaporations were carried out at room temperature andat 60 ◦C. Thus Skopintsev’s “volatile fraction” consists of those compoundsthat are volatile from acidified solution taken to dryness at 60 ◦C but not at20 ◦C. This fraction was found to be between 10 and 15% of the total organiccarbon. He also noted a 15% difference in measured organic carbon with hisdry combustion method when samples were dried at different temperaturesand concluded that this difference was due to the loss of volatiles.

The volatile fraction as defined by the various wet oxidation methods andmost of the direct injection methods would be that fraction removed by acidi-fication and purging with inert gas at room temperature. In the freeze-dryingmethod of Gordon and Sutcliffe [29] the volatile fraction is that fraction lostby sublimation in vacuo. There have been no actual determinations of theselosses, and for the most part Skopintsev’s numbers were accepted as valid forall of these methods, largely because they are the only numbers available.

MacKinnon [91] and Wangersky [180] have made direct determinations ofthe volatile fractions from a variety of depths and stations in the North Atlantic.The volatile fraction as defined by MacKinnon’s method is that fraction whichcan be removed from solution by purging with an inert gas at 80 ◦C and a pHof 8 for 10–12 hours, then at 65 ◦C for a further 10–12 hours. The inert gasstream is flushed through an ice-packed condenser to remove water, then intoa trap packed with Tenax GC followed by a U-shaped stainless steel cold trapheld at –78 ◦C.

Gershey et al. [58] have pointed out that persulfate and photo-oxidationprocedures will determine only that portion of the volatile organics not lostduring the removal of inorganic carbonate [30,79,92,181]. Loss of the volatilefraction may be reduced by use of a modified decarbonation procedure such asone based on diffusion [98]. Dry combustion techniques that use freeze-dryingor evaporation will result in the complete loss of the volatile fraction [72, 79,92, 93].

The loss of volatile organics will only be a problem in areas where the volatilecomponent is high. In open ocean areas volatiles should be a small fraction(2–6%) of the total organic carbon [91]. Under strongly reducing conditions,


many of the end-products of microbial metabolism are volatile in nature.Deuser [182] noted that his wet oxidation analysis of the dissolved organiccarbon of water from the reducing part of the Black Sea gave systematicallyhigher values than those of dry combustion. He concluded that it was due tosubstantial amountsof volatilematerial in thewaterproducedas a consequenceof anaerobic microbial metabolism.

In further work to that discussed above, MacKinnon [183] has discussed indetail a method for the measurement of the volatile fraction of total organiccarbon in seawater.

In this method volatile organic matter in seawater is concentrated on a TenaxGCsolidadsorbent trapanddry-ice trap in series.The trappedorganicmaterialis then desorbed and oxidised to carbon dioxide, which is measured witha non-dispersive infrared analyser. A dynamic headspace method was usedfor the extraction with the assistance of nitrogen purging. Dynamic headspaceanalysis [184] is an efficient extraction procedure. The efficiency of extraction

Table 11.3. Averaged volatile organic carbon (VOC) concentrations and VOC:total organiccarbon (VOC:TOC) ratios from different areas [92]

Area (timeof sampling) Depth n AverageVOC con-centrations(µg C/l)

Range(µg C/l)

AverageVOC:TOCratio (%)

Range< (%)

Gulf of St. Lawrence 0–10 9 31.8 ± 8.4 21.7–51.7 2.7 ± 0.9 1.9–4.8(11/75) 10–50 6 36.0 ± 8.9 22.3–47.0 3.0 ± 0.6 2.0–3.7

50–100 5 29.8 ± 1.7 27.5–32.2 2.7 ± 0.4 2.2–3.1100–250 9 30.1 ± 6.5 22.8–38.8 3.1 ± 0.7 2.2–4.1

Scotian shelf & slope 0–10 26 40.9 + 9.1 23.9–60.6 3.3 ± 0.9 1.8–5.9(5/74) (8/75) (3/76) 10–25 16 37.8 ± 8.1 24.0–53.0 3.4 ± 0.8 1.9–5.2

25–100 17 33.9 ± 9.9 24.0–68.5 3.5 ± 0.9 1.8–6.2100–250 17 35.5 + 3.7 28.3–41.2 4.1 ± 0.5 3.5–4.9250–750 12 33.0 ± 6.4 22.4–46.4 4.5 ± 0.9 3.0–6.1750–1500 5 26.2 ± 2.4 23.2–29.7 3.7 + 0.3 3.4–4.0

Central & northwestern 0–10 31 30.4 ± 9.7 17.6–52.4 3.1 + 1.1 1.7–5.5Atlantic (10/74) (2/75) 10–25 14 27.8 + 6.2 18.8–36.7 2.8 + 0.7 1.7–3.2

25–100 15 31.9 ± 9.8 22.6–52.2 3.4 ± 1.1 2.3–5.8100–250 35 27.7 ± 6.5 17.0–52.3 3.3 ± 0.9 2.0–5.8250–750 22 25.1 + 4.4 19.6–34.0 3.5 ± 0.7 2.5–5.1750–1500 15 26.1 ± 4.6 20.2–34.7 3.6 ± 0.7 2.7–4.7

1500–3000 15 24.4 ± 2.5 19.2–29.3 3.4 + 0.4 2.5–4.33000–5000 5 26.5 ± 1.7 25.0–29.1 3.8 ± 0.5 3.4–4.7

Northwest arm HalifaxHarbour (4/76)

0–10 37 32.6 ± 7.8 20.4–49.4 1.9 ± 0.4 1.2–2.9

St. Margaret’s Bay, NovaScotia (4/76)

0–40 28 30.9 + 6.2 19.8–45.5 2.3 ± 0.5 1.4–3.4


can be increased by using a purging gas and increasing the temperature of thesample to 60–80 ◦C [173]. The efficiency and properties of Tenax GC have beenexamined [177, 185–187].

Theefficiencyof extractionandanalysiswasdetermined fordifferent classesof volatile organic materials. Acetone was used as a test material to evaluateefficiency for the extraction, concentration, and analysis procedures. Whileacetone has a high vapour pressure, it is difficult to extract efficiently fromwater because of its high solubility. Almost complete recovery of acetone fromseawater was obtained. Over the concentration range 10–200 µg C per litre,high recoveries (100±20%)of acetoneadded topre-extracted seawater sampleswere calculated. For other compounds tested, high recoveries were noted formaterials of low solubility and high vapour pressure (hydrocarbons, ethers,esters, aldehydes, ketones, nitriles), while poorer recoveries were observed forthe more polar organic compounds (alcohols, amines, acids).

Table11.3presents some typical resultsobtainedby thisprocedureonseawa-ter samples. MacKinnon [92] concluded that since the volatile organic carboncontents of normal (i.e., unpolluted) seawaters are small, the effect of completeor partial loss of volatile organic components during the determination of to-tal organic carbon in most ocean areas (except highly reducing environments)with either the wet and direct injection methods or dry oxidation methodsshould be small (about 5%), and within the precision of these methods.

11.7.5Chemical Oxygen Demand

The chemical method for the determination of the chemical oxygen demandof non-saline waters involves oxidation of the organic matter with an excessof standard acidic potassium dichromate in the presence of silver sulfate cat-alyst followed by estimation of unused dichromate by titration with ferrousammonium sulfate. Unfortunately, in this method, the high concentrations ofsodium chloride present in sea water react with potassium dichromate pro-ducing chlorine:

16Cl– + Cr2O2+7 + 14H → 2Cr3+ + 3Cl2 + 7H2O

Consequently the consumption of dichromate is many times higher than thatdue to organic material in the sample. To complicate matters, any amines inthe sample consume and release chlorine in a cyclic process, leading to highchemical oxygen demands.

RNH2 → NH2 + CO2 + H2O

NH+4 + 3Cl2 → NCl3 + 3Cl– + 4H+

2NCl3 → N2 + 3Cl2


Also, the addition of silver sulfate causes precipitation of silver chloride, whichin the presence of organic compounds is neither completely nor reproduciblyoxidised. This method, whilst being applicable to estuarine waters of relativelylow chloride content, would present difficulties when applied to highly salineestuarine and sea waters of low organic content.

This method, whilst being applicable to estuarine waters of relatively lowchloride content, would present difficulties when applied to highly saline sea-waters of low organic content. Zeitz [188] gives details of a method for thedetermination of chemical and oxygen demand, by potassium dichromate.Its advantages over previously published methods are that greater accuracyis achieved in recording even very small chemical oxygen demand measure-ments when there is a high chloride concentration, and the use of mercurycompounds as buffering agents is obviated. It is claimed that the process mustbe carried out in a closed vacuum flask. The concentration of chlorine in thegaseous phase is so slight that it can be disregarded when chemical oxygendemand values are being determined.

Wagner and Ruck [189, 190] propose various methods for overcominginterference by chloride in the determination of chemical oxygen demand.These include the quantitative oxidation of chloride ions to chlorine anda corresponding correction of the result of the chemical oxygen demanddetermination; use of mercuric sulfate as a sequestering agent for chlorideions, for which the limits of validity are discussed; and elimination of chlo-ride ions prior to the oxidation step by conversion to hydrogen chlorideand its diffusion into a closed chamber to reduce losses of volatile con-stituents.

Because of the invalidity of the classical procedure, several workers haveattempted to devise a method that is free from interference by chloride. Chlo-ride interference can be eliminated by preventing the concurrent oxidationof organic material and chloride. This can be effected in two ways – ei-ther by leaving the chloride in the test mixture but preventing its oxida-tion, or by removing the chloride prior to the chemical oxygen demandtest.

Both ways have been used in previously reported attempts to remove chlo-ride interference. Three methods have been used in attempts to prevent chlo-ride oxidation: masking with mercury (II) [191,192]; precipitation of chlorideusing silver (I) [188]; or altering oxidation conditions [193].

Two methods have been used for removal of chloride, the chloride beingremoved as chlorine [194, 195] or as hydrogen chloride [196].

Baumann [197] collected the chlorine produced in the reaction in excesspotassium iodide solution and back titrated against standard sodium thiosul-fate to obtain the necessary correction for chloride.

Wagner and Ruck [190] carried out experiments to test the reliability oftwo standard procedures (DIN 38409 Parts 41/2 and 42) for the removal ofchloride ions prior to chemical oxygen demand determinations on chloride-


containing samples. Acetic acid was used as a model substance on accountof its volatility, and the use of special absorbers for hydrochloric acid vapourgenerated by the addition of sulfuric acid was tested under various condi-tions. In no case was there any appreciable loss of acetic acid, but in somecases a yellow coloration was observed in the solution coupled with an in-crease in the measured chemical oxygen demand value. This was traced tothe presence of nitrate ions which were reduced to nitrite ions by chloridein acid solution, the resulting nitrite ions being reoxidised during the sub-sequent stages and hence contributing to an inflated value for the chemicaloxygen demand. As a preventive measure in the case of samples containingnitrate ions the hydrogen chloride diffusion can be carried out using weakersulfuric acid solutions and, if necessary, using a longer reaction time in theanalysis.

Southway and Bark [198] reported their findings on their investigations intothree processes for removing chloride from the test solution and suggested thatthe method (3) may be of some value:

1. Removal by precipitation as silver chloride with subsequent filtration, sothat a negligible amount of organic matter is coprecipitated with the silverchloride.

2. Removal as chlorine in an oxidation stage, mild enough to prevent anysignificant oxidation of easily oxidised organic matter.

3. Removal by a ligand exchange process using the silver I or the mercury IIform of poly vinyl pyridine [199].

Lloyd [200] has given details of a method for determining the chemical oxy-gen demand of saline samples which uses digestion of the sample at 150 ◦Cin a glass stoppered flask with silver nitrate to suppress interference by chlo-ride. Within-batch relative standard deviation ranged from 2.2% at levels of60 mg/l of chemical oxygen demand to 0.8% at 380 mg/l of chemical oxygendemand (4 degrees of freedom). Total relative standard deviation in analysisof a 300 mg/l chemical oxygen demand potassium hydrogen phthalate solutionwas 1.0% (8 degrees of freedom) These did not differ significantly (p = 0.5)from data obtained using the standard procedure or the silver nitrate refluxchemical oxygen demand procedure.

Standard solutions of potassium hydrogen phthalate, spiked with chloride,were also analysed. These results are compared with those given for the stan-dard procedure in Table 11.4.

Thompson et al. [201] have described a simple method for minimising,but not eliminating, the effect of chloride on the determination of chemicaloxygen demand without the use of mercury salts. This method minimisedinterference from the potassium dichromate oxidation of chloride to chlorineby adding a small amount of chromium (III) to the sample digest prior toheating. Chromium (III) was thought to complex any free chloride ion in thesample solution. The method was particularly useful at low chemical oxygen


Table11.4. Analyses of chloride-spiked potassium hydrogen phthalate solutions by standardand sealed-tube procedures (1 ml 25% m/v silver nitrate solution) [200]

Expected COD(mg/l)

Observed COD at given chloride level (mg/l Cl)

0 500 1000 1500 2000

0 – 15 (10) 19 (23) 21 (41) 32 (59)100 102 112 (103) 115 (209) 121 (109) 128 (121)200 199 203 (206) 209 (214) 221 (215) 217 (228)300 295 300 (308) 297 (308) 311 (311) 316 (318)400 391 401 (404) 403 (404) 405 (412) 407 (419)

Results are the means of two determinations. Values in parentheses are means ofresults obtained using the standard sealed-tube procedure.

demandconcentrations, chlorideeffects significantlydecreasingwith increasesin chemical oxygen demand.

In the author’s experience, no truly satisfactory method yet exists for thedetermination of chemical oxidation demand in highly saline samples suchas seawater. With reservations, several of the above methods are, however,applicable to estuarine waters of low salinity.

11.7.6Biochemical Demand

It has been reported that whilst the dilution bottle method for biochemicaloxygen demand yields satisfactory results on fresh and low saline waters,a discrepancy exists when the test is performed on waters containing elevatedlevels of sodium chloride and other salts.

Kessick and Manchen [202], using a manometric BOD apparatus, found thatthe solublewasteportion ina saltwater carrierwasmetabolisedat the same rateby themicrobialpopulation indomestic sewage.However, the insolubleportionproduceddifferent rates,with thatofdomestic sewage in freshwaterbeing threetimesgreater than thatofdomestic sewage inasaltwater carrier.Manchen[203]reported that only the degradationof suspended organic material was inhibitedby salt and that the dissolved organic fraction of the BOD was unaffected.Davis et al. [204], using hypersaline industrial wastes with salt tolerant bacteriaspecies as seed, found that the BOD varied according to the salt concentrations.Increases in BOD resulted when the salt level was decreased. Ten-day BODson salt water wastes were 3–6 times higher with standard dilution water thanwith a dilution water having the same salinity as the waste.

The rate of biological oxidation may vary considerably under differenttest conditions. Gotaas [205] reported that waste waters containing less than10 000 mg/l chloride diluted with fresh water had a rate of biological oxidation


value greater than that of fresh water waste. The rate of biological oxidationincreased to the point equivalent to 50% of that of seawater. When the 50%level was exceeded, the rate of biological oxidation decreased, until in 100%seawater, it was lower than in fresh water. This led to the hypothesis that a lowsalinity may stimulate microbial activity and result in increased BOD values.Seymour [206] also found that an increase in salinity resulted in a decreaseof BOD values. Degradation rate decreased as the salinity level increased, yetonly the particulate fraction seemed to be affected. Degradation of the solubleportion was not affected as adversely by salt content.

Lysis of bacteria will occur in a growth medium when sodium chlorideis added. It will also occur when salt-acclimated bacteria are placed in freshwater. If the change in salt concentration is less than 10 000 mg/l the degree oflysis is negligible, but if greater the lysis can be extensive. Cellular constituentsreleased following lysis are metabolised by the remaining microbial populationin preference to the substrate present, thus yielding erratic results [207, 208].

Davies et al. [209] set out to quantify the effects of fresh and saline dilutionwaters on the BOD of salt water wastes and to quantitatively establish the roleof bacterial seed numbers and species associated with changes in BOD results.Sewagebacteria and salt-tolerantbacteriawere evaluated todetermine their ca-pability to produce the same BOD values at various salt concentrations. Whenvariations occurred in the BOD values, Davies et al. [209] attempted to identifythe reasons for such variations. They employed conventional and manomet-ric BOD methods. Standard organic solutions and an industrial waste weretested with sewage seed and known species of salt-tolerant bacteria using stan-dard and hypersaline dilution water at three salt concentrations. SignificantBOD differences were found when saline wastes were diluted with standard(non-saline) BOD dilution water. Bacterial populations to genera were moni-tored and it was shown that equivalent numbers of bacteria did not have thesame capability to degrade a given amount of waste with increases in salt con-centrations to the 3% level. Seeding of hypersaline waste waters with knownsalt-tolerant species is recommended for consistent BOD results.

The BOD values obtained in sewage seed organic standards showed a sig-nificant trend: as salt concentration increased, BOD values decreased. Changeswere significant at the 0.05 level. Dilution of sewage-seeded saline organicstandards with standard dilution water resulted in BOD values higher thanthose of corresponding non-saline organic standards, due to increases in bac-teria populations and increased organic removal (BOD values) in the presenceof low levels of salt. Bacteria populations of sewage seed correlated with cor-responding BOD values and concentrations of salt in organic standard solu-tions showed that the addition of 1% and 3% salt resulted in decreased initialbacteria population, decreased growth rates, and a decreased ability of anequivalent number of bacteria to degrade an equivalent amount of organicmaterial.


Each salt-tolerant bacteria species yielded comparable BOD values for eachorganic standard conducted at three salt concentrations. Population increasedand the ratesofbiological oxidationwere similar at the three salt concentrationsand when using either the standard dilution water or the saline dilution water.

Since sewage seed does not compare favourably with salt-tolerant bacteriain determining the BOD of saline waste waters, a salt-tolerant bacteria seedshould be used to ensure an accurate and reproducible BOD value hypersalinewaste waters are being tested.

11.8Oxygen Isotopic Ratios

Baker et al. [210] described a technique for the determination of 18O/16O and17O/16O isotopic ratios in liquid samples.

References

1. Ball JW, Thompson JM, Jenne EA (1978) Anal Chim Acta 98:672. Chen L, Lin J, Yu J, Wu S (1993) Haiyang Xuebao (Zhongwenban) 15:1423. Nishioka H, Sai N, Kumagai T, Nagahiro T, Uesugi K (1992) Nippon Kaisui Gakkaishi

46:3914. Schnepfe MM (1972) Anal Chim Acta 58:835. Sugawara D (1956) Anal Abst 93:11696. Matthews AD, Riley JP (1970) Anal Chim Acta 51:2957. Tsunogai S (1971) Anal Chim Acta 55:4448. Kesari R, Rastogi A, Gupta VK (1998) Chem Anal (Warsaw) 43:2019. Astrani U (1973) Analysis of waters for nitrogen. PB report no. 238056/6GA. National

Technical Information Service, Springfield, VA, USA10. Yablokava OG (1974) Okeanologiya 14:111311. Mertens J, Van Den Winkel P, Massart DL (1975) Int J Environ Anal Chem 47:2512. Stevens RJ (1976) Water Res 10:17113. Adamski JM (1976) Anal Chem 48:119414. D’Elia CF, Steudler PA, Corwin N (1977) Limnol Oceanogr 22:76015. Armstrong FAJ, Williams PM, Strickland JDH (1966) Nature 211:48116. Tenny AM (1966) Automated Anal Chem 38:58017. Manny BA, Miller MC, Wetzel RG (1971) Limnol Oceanogr 16:7118. Bruegmann L, Wilde A (1975) Acta Hydroch Hydrob 3:20319. Holm-Hansen O, Strickland JDH, Williams PM (1966) Limnol Oceanogr 11:54820. Butler EI (1979) Estuar Coast Mar Sci 8:19521. Butler E, Knox S, Liddicoat MI (1979) J Mar Bio Assoc UK 59:23922. Strickland JDH, Parsons TR (1972) A practical handbook of seawater analysis. Bulletin

167. Fisheries Research Board, Canada, p 31023. Armstrong FAJ, Tibbitts S (1968) J Mar Bio Assoc UK 48:14324. Koroleff F (1969) Determination of total nitrogen in natural waters by means of diges-

tion. International Council for the Exploration of the Sea, Hydrography Committee,CM/C:8, p 4

25. Koroleff F (1976) In: Grasshoff K (ed), Methods of seawater analysis. Verlag Chemie,New York, p 116

References 499

26. Shepherd RJ, Davies IM (1981) Anal Chim Acta 130:5527. D’Elia CF, Stendler PA, Corwin N (1977) Limnol Oceanogr 22:76028. Nydahl F (1978) Water Res 12:112329. Gordon DC Jr, Sutcliffe WH Jr (1973) Mar Chem 1:23130. Collins KJ, Williams PJ, Le B (1977) Mar Chem 5:12331. Suzuki A, Satah M, Kamiya M, Ogawa S (1985) Mizushori Gijutsu 26:92132. Nakamura E, Namiki H (1987) Kogyo Yosui 341:1233. Frankovitch TA, Jones RD (1998) Mar Chem 60:22734. Duursma EK (1960) Neth J Mar Res 1:14835. Harvey HW (1948) J Mar Bio Assoc 27:33736. Armstrong FAJ, Tibbitts S (1968) J Mar Bio Assoc 48:14337. Grasshoff K (1966) Fresen Z Anal Chem 220:8938. Bikbulatov EW (1974) Gidrokhim Mater 60:17439. Mueller H, Bandaranayake WM (1983) Mar Chem 12:5940. Cembella AD, Antia NJ, Taylor FJP (1986) Water Res 20:119741. Ormaza-Gonzales FI, Staham PJ (1996) Water Res 30:273942. Abe K, Wanatabe Y (1992) J Oceanogr 48:28343. Taylor DL, Zeitlin H (1973) Anal Chim Acta 64:13944. Wangersky PJ (1975) Measurement of organic carbon in seawater. In: Analytical Meth-

ods in Oceanography (ed Gibb TRP) American Chemical Society Washington DC, p148–162

45. Ogawa H (1993) Shigen Kankyo 2:1346. Menzel DW, Vaccaro RF (1964) Limnol Oceanogr 9:13847. Strickland JDH, Parsons TR (1968) Bulletin Fisheries Research Board of Canada, pp

167–31148. Sharp JH (1973) Mar Chem 1:21149. Skopintsev BA (1960) Marine Hydrophysics Institute 19:150. Takahashi Y (1976) Ultra low level TOC analysis of potable waters. Presented at the

water quality technology conference. American Water Works Association, San Diego,CA, USA

51. Van Hall CE, Safranko J, Stenger H (1963) Anal Chem 35:31552. Axt G (1969) Vom Wasser 36:32853. Solomen K (1979) Limnol Oceanogr 24:17754. Glor R, Leidner H (1979) Anal Chem 51:64555. Beattie J, Bricker C, Garvin D (1961) Anal Chem 33:189056. Woelfel P, Sontheimer H (1974) Vom Wasser 43:31557. Ehrhardt M (1969) Deep-Sea Res 16:39358. Gershey RM, MacKinnon MD, Williams PJ, Le B, Moore RM (1979) Mar Chem 7:28959. Wangersky PJ (1981) Trends Anal Chem 1:2860. Armstrong FAJ, Boalch GT (1961) J Mar Bio Assoc 41:59161. Bauer D, Brun-Cottan JC, Saliot A (1971) Canadian Oceanography 23:84162. Hanya T, Ogura N (1964) Columbo U, Hobson GD (eds) Application of ultraviolet

spectroscopy to the examination of dissolved organic substances in water. In: Advancesin organic geochemistry. Pergamon Press, Oxford, pp 1–10

63. Ogura N, Hanya T (1966) Nature 212:75864. Ogura N, Hanya T (1967) Limnol Oceanogr 1:9165. Lewis WM Jr, Tyburczy JA (1974) Archives of Hydrobiology 74:866. Mattson JS, Smith CA, Jones TT, Gerchakov SM (1974) Limnol Oceanogr 19:53067. Wheeler JR (1977) Limnol Oceanogr 22:57368. Kulkarni RN (1975) Water Air Soil Poll 5:231


69. Biggs R, Schofield JW, Gorton PAJ (1976) Water Pollut Cont Fed 47:270. Maciolek JA (1962) USF and WS Research Report 60:171. Camps JM, Arias E (1963) Invest Pesq 23:12572. Oppenheimer CH, Corcoran EF, Van Arman J (1963) Limnol Oceanogr 8:48773. Krey J, Szekielda KH (1965) Fresen Z Anal Chem 207:33874. Duursma EK (1960) Netherlands J Mar Res 1:175. Williams PJ, Le B (1969) Limnol Oceanogr 14:29276. Boehme H (1975) Acta Hydroch Hydrobiol 3:32777. Baldwin JM, McAtee RE (1974) Microchem J 19:17978. Goulden PD, Brooksbank P (1975) Anal Chem 47:194379. Wilson RF (1961) Limnol Oceanogr 6:25980. Martinie GC, Schilt AA (1976) Anal Chem 48:7081. Williams PH (1969) Limnol Oceanogr 14:29782. Hannaker P, Buchanan SS (1983) Anal Chem 55:192283. Gordon MA (1914) J Phys Chem 18:5584. Williams PH (1969) Anal Abst 17:361385. Woelfel P, Schuster W, Sontheimer H (1975) Z Wass Abswass Fors 8:14386. Schreurs W (1978) Hydrobiol Bull 12:13787. House DA (1962) Chem Rev 62:18888. Moore RM (1977) Trace metals dissolved organic matter and their association in

natural waters. PhD Dissertation. University of Southampton89. Montgomery HAC, Thorn NS (1962) Analyst 87:68790. Skopintsev BA, Timofeyeva SN (1962) Marine Hydrophysics Institute 25:11091. MacKinnon MD (1977) The analysis of total organic carbon in seawater. (a) Develop-

ment of methods for the quantification of TOC. (b) Measurement and examination ofthe volatile fraction of the TOC. PhD Dissertation. Dalhousie University

92. MacKinnon MD (1978) Mar Chem 7:1793. Skopintsev BA (1966) Oceanography 6:36194. Williams PJ, Le B (1975) Determination of organic components. In: Riley JP, Skirrow

G (eds) Chemical oceanography, vol. 3. Academic Press, New York, pp 443–47795. Skopintsev BA (1976) Oceanology 16:63096. WangerskyPJ,ZikaRG(1978)Theanalysisoforganiccompounds in seawater.Report3,

NRCC 16566, Marine Analytical Chemistry Standards Program97. Van Hall CE, Stenger VA (1964) Water Sewage 111:26698. Van Hall CE, Barth D, Stenger VA (1965) Anal Chem 37:76999. Van Hall CE, Stenger VA (1967) Anal Chem 39:503

100. Van Hall CE, Safranko J, Stenger VA (1963) Anal Chem 35:319101. Golterman HL (1969) Methods for chemical analysis of freshwater. Blackwell, Oxford,

pp 133–145102. Water Pollution Research Laboratory (1972) Notes on water pollution, no. 59. HMSO,

London103. Salvatella Battlori N, Ribas Soler F, Oromi Durich J (1974) Doc Invest Hidrol 17:303104. Stenger V, Van Hall CE (1969) Anal Chem 39:206105. Eggertsen FT, Stross FH (1972) Anal Chem 44:709106. Bartz AM (1968) SPE Transactions 7:279107. Hiser LL, Tarazi DS, Boldt CA, Saenz O (1970) Am Lab 2:44108. Ton N, Takahashi Y (1985) Int Lab September:49109. Nelson KH, Lysyj I (1968) Environ Sci Tech 2:61110. Dobbs RA, Wise RH, Dean RB (1967) Anal Chem 39:1255111. Cropper FR, Heinekey DM, Westwell A (1967) Analyst 92:436

References 501

112. Cropper FR, Heinekey DM, Westwell A (1967) Analyst 92:443113. Croll BT (1972) Chem Ind London 386114. Merks AGA, Vlasbom AG (1983) Hydrobiol Bull 17:53115. Olrichs SHS (1982) Projektur 71:211116. Elgershuizen JHBW (1981) Meitingen von DOC Gehahen, Notitie, DDMI 81-278117. Games LM, Hayes JM (1976) Anal Chem 48:130118. Mitchell DG, Aldous KM, Canelli E (1977) Anal Chem 49:1235119. Mills GL, McFadden E, Quinn JG (1987) Mar Chem 20:313120. Neihof R, Loeb G (1974) J Mar Res 32:5121. Chave KE (1965) Science 148:1723122. Meyers PA, Quinn JG (1971) Limnol Oceanogr 16:992123. Meyers PA, Quinn JG (1973) Geochim Cosmochim Acta 37:1745124. Jeffrey LM, Hood DW (1958) J Mar Res 17:247125. Williams PM, Zirino A (1964) Nature 204:462126. Chapman G, Rae AC (1967) Nature 214:627127. Keeling CD, Whorf TP (1994) ORNL/CDIAC 65:16128. Sarmiento JL (1993) Nature 365:697129. Tans PP, Fung IY, Takahashi T (1990) Science 247:1431130. Sarmiento JL, Orr JC, Siegenthaler UL (1992) J Geophys Res C97:3621131. Johnson KM, Williams PJ, Le B, Brandstrom L, Sieburth JM (1987) Mar Chem 21:117132. Johnson KM, Wills KD, Butler DB, Johnson WK, Wong CS (1993) Mar Chem 44:167133. Goyet C, Hacker SD (1992) Mar Chem 38:37134. DOE (1994) Dickson AG, Goyet C (eds) ORNL/CDIAC 74135. Sugimura Y, Suzuki Y (1988) Mar Chem 24:105136. Wiebinga CJW (1999) PhD Dissertation. University of Groningen137. Godec R, O’Neill K, Hutte R (1984) Ultrapure Water 17138. Bakker DCE, de Baar HJW, Bathmann U (1997) Deep-Sea Res 44:91139. Takahashi T, Olafsson J, Goddard JG, Chipman DW, Sutherland SC (1993) Global

Biogeochem Cy 7:843140. Hoppema JMJ, Fahrbach E, Schröder M, Wisotzki A, de Baar HJW (1995) Mar Chem

51:177141. Hall POJ, Aller RC (1992) Limnol Oceanogr 37:1113142. Stoll MHC, Bakker K, Nobbe GH, Haese RR (2001) Anal Chem 73:4111143. Fox D, Isaacs JD, Corcoran EF (1952) J Mar Res 11:29144. Kuck JA, Berry JW, Andreatch AJ, Lentz PA (1962) Anal Chem 34:403145. Ernst W (1975) Inst Meeresforsch Bremerh 15:269146. Szekielda KH, Krey J (1965) Mikrochim Acta 149147. Hagell GT, Pocklington R (1973) A seagoing system for the measurement of particulate

carbon, hydrogen and nitrogen. Report Ser./BI-R-73-14, 1–18148. Telek G, Marshall N (1974) Mar Biol 24:219149. Ehrhardt M (1976) Determination of particulate organic carbon and nitrogen. In:

Grasshoff K (ed) Methods of seawater analysis. Verlag Chemie, New York, pp 215–220150. Ehrhardt M (1976) The automatic determination of dissolved organic carbon. In:

Grasshoff K (ed) Methods of seawater analysis. Verlag Chemie, New York, pp 289–297151. Hirota J, Szyper JP (1975) Limnol Oceanogr 20:896152. Mills JG, Quinn GL (1979) Chemical Geology 25:155153. Kerr RA (1977) The isolation and partial characterization of dissolved organic matter

in seawater. PhD Dissertation. University of Rhode Island154. Weliky K, Suess E, Ungerer CA, Muller PJ, Fischer K (1983) Limnol Oceanogr 28:1252155. Salonen K (1979) Hydrobiologia 67:29


156. Giger W (1977) Mar Chem 5:429157. Schwarzenbach RP, Bromond RH, Gschwend PM, Zafiriuo OC (1978) Org Geochem

1:93158. Armstrong FA, Boalch GT (1960) Nature 185:761159. Koyama T (1962) Journal of the Oceanographic Society of Japan 20:563160. Lovelock JE, Maggs RJ, Rasmussen RA (1972) Nature 237:452161. Wangersky PJ (1976) Deep-Sea Res 23:457162. Wilson DF, Swinnerton JW, Lamontagne RA (1970) Science 168:1577163. Dowty BL, Green LE, Lasteter JL (1976) Anal Chem 46:946164. Zika RG (1977) An investigation in marine photochemistry. PhD Dissertation. Dal-

housie University165. Goldberg ED (1971) Atmosphere transport. In: Hoo DW (ed) Impingement of man on

the oceans. Wiley-Interscience, New York, pp 75–88166. Swinnerton JW, Lamontagne RA (1971) Environ Sci Tech 8:657167. Lamar WL, Goerlitz DF (1966) Organic acids in naturally colored surface waters.

Geological Survey Water Supply Paper 1817-A168. Ryabov AK, Nabivanets BI, Litvinenko SZ (1972) Hydrobiological Journal 8:97169. Kamata E (1966) Bulletin of the Chemical Society of Japan 39:1227170. Bassette R, Ward G (1969) Microchem J 14:471171. Corwin JF (1970) Volatile organic matter in seawater. In: Hood DW (ed) Organic matter

in natural waters, vol. 1. University of Alaska Press, pp 169–180172. Hurst RE (1974) Analyst 99:302173. Mieure JP, Dietrich MW (1973) J Chromatogr Sci 11:559174. Novak J, Zluticky J, Kubelka V, Mostecky J (1973) J Chromatogr 76:45175. Grob K (1973) J Chromatogr 84:255176. Zurcher F, Giger W (1976) Vom Wasser 47:37177. Kuo PRK, Chian ESK, DeWalle FB, Kim JH (1977) Anal Chem 49:1023178. Sauer TC Jr, Sackett WM, Jeffrey LM (1978) Mar Chem 7:1179. Harris LE, Budde WL, Eichelberger JW (1974) Anal Chem 46:1912180. Wangersky PJ (1978) Production of dissolved organic matter. In: Kinne O (ed) Marine

ecology, vol. 4. Wiley, New York181. Duursma EK (1961) J Sea Res 1:1182. Deuser WS (1971) Deep-Sea Res 18:995183. MacKinnon MD (1979) Mar Chem 8:143184. Grob K, Grob K Jr, Grob G (1975) J Chromatogr 106:299185. Zlatkis A, Lichtenstein HA, Tichbee A (1973) Chromatographia 6:67186. Russell JW (1975) Environ Sci Tech 9:1175187. Daemen JM, Dankelman HW, Hendriks ME (1975) J Chromatogr Sci 13:79188. Zeitz U (1976) Gas-u-Wassertach (Wasser Abwasser) 117:181189. Wagner R, Ruck W (1981) Z Wasser Abwass For 14:145190. Wagner R, Ruck W (1982) Z Wasser Abwass For 15:287191. Dobbs RA, Williams RT (1963) Anal Chem 35:1064192. Cripps JM, Jenkins D (1964) Journal of the Water Pollut Cont Fed 36:1240193. Ryding SO, Forsberg A (1977) Water Res 11:801194. Cameron WM, Moore TB (1957) Analyst 82:677195. Hoffmann HJ (1979) Labor Prax (Warsberg) 3:38196. Frankland E, Armstrong HE (1868) Journal of the Chemical Society 21:77197. Baumann FJ (1974) Anal Chem 46:1336198. Southway C, Bark LS (1981) Anal Proc 18:7199. Heinerich F (1961) Nature 189:1001

References 503

200. Lloyd A (1982) Analyst 107:1316201. Thompson KC, Mendham D, Best D, De Casseres KE (1986) Analyst 111:483202. Kessick MA, Manchen KL (1976) Journal of the Water Pollut Cont Fed 48:2131203. Manchen LK (1974) Unpublished dissertation, Rice University204. Davis EM, Petros JK, Powers EL (1977) Industrial Wastes 23:22205. Gotaas HB (1949) Sewage Works J 21:818206. Seymour MP (1977) The biodegradation of particulate organic matter in saltwater

media. Dissertation, Rice University207. Kincannon DF, Gaudy AF (1966) Biotechnical Engineering 8:371208. Kincannon DF, Gaudy A (1966) Journal of the Water Pollut Cont Fed 38:1148209. Davies EM, Bishop JR, Guthrie RK, Forthofer R (1978) Water Res 12:917210. Baker L, Franchi IA, Maynard J, Wright IP, Pillinger CT (2002) Anal Chem 74:1165

Subject Index

acetate, determination of 39acidity, determination of 21acrylamide, determination of 400, 401acrylate, determination of 39actinium, determination of 112adsorption techniques 354aldehydes, determination of 380, 381aliphatic chloro-compounds,

determination of 390–392aliphatichydrocarbons,determinationof

366–368alkalinitydetermination 39–41preservation 21alkyl lauryl sulfide, determination

of 401aluminium– determination of 112–114, 247,

271, 322– sampling 2americium, determination of 336amines, determination of 398ammonium– determination of 114–118, 313– preservation 21amperometry, determination of anions

103–108anions, pre-concentration 91, 92anionic surfactants, determination

of 22, 386–388anodic stripping voltammetry,

determination of– cations 114, 123, 134, 166, 172, 210,

217, 248, 316, 322– organometallic compounds 6, 44antimonate, determination of 91antimony– determination of 119, 229, 233, 235,

244, 247, 253, 262, 323

– sampling of 2, 3aromatic hydrocarbons, determination

of 368, 369arsenate, determination of 41, 88,

89, 91arsenic– determination of 120–124, 219, 232,

238, 241–247, 262, 313, 314, 323– preservation 24– sampling 2arsenite, determination of 41–88,

89, 91atomic absorption spectrometry– of anions 48, 49, 56, 65, 101– of cations 101, 114, 119, 121, 124, 125,

126, 128, 129, 135, 138, 148, 154, 163,166, 174, 203, 205, 207, 213, 216, 220,223–239, 314–318, 323

– of organic compounds 379– of organometallic compounds 444,

450, 455atomic emission spectrometry,

determination of organometalliccompounds 459

atomic fluorescence spectroscopy,determination of cations 150, 168

azarines, determination of 400azine herbicides, determination of 411

barium, determination of 124, 125, 262,265, 266,313, 314

benzoate, determination of 41beryllium, determination of 125, 247biochemical oxygen demand

determination of 21, 496bismuth, determination of 126, 127,

229, 233, 244, 253Blumer sampler 5Bodeja-Bodman bottle 6

506 Subject Index

borate, determination of 42boron, determination of 127, 128, 467bromate, determination of 42–45bromide, determination of 45, 46,

89, 90bromine, determination of 266bromo-organic compounds,

0 determination of 49, 50butyrate, determination of 41

cadmium– determination of 14, 128–134, 220,

223–225, 229–232, 241–270, 315,320–323, 365

– preservation 22, 23caesium,– determination of 136, 270– preservation of 24, 25calcium, determination of 136–139,

219, 262, 316carbohydrates, determination of

382, 385carbon dioxide, determination of 108,

109carbon functions, determination of

472–493carboxylic acids, determination of

377–379cathodic stripping voltammetry– determination of anions 63, 69, 72– determination of cations 151, 191,192,

202, 211, 214, 259, 322cationic surfactants, determination

of 386cations, preconcentration 288, 374cerium, determination of 136, 194, 196,

266, 267chelators, determination of 415chemical oxygendemand,determination

of 21, 490–495chemiluminescence of– anions 68– cations 150, 166, 219chloride, determination of 47, 48chlorodioxines, determination of 393chloroinsecticides, determination of

403–409chlorophyll, determination of 20chlorothalonil, determination of 412chromate, determination of 48, 53, 91

chromium, determination of 139–145,224, 232, 241–251, 261, 265, 267, 270,321, 322

chronopotentiometry– determination of anions 48– determination of cations 260coastal water– determination of anions 99–102cobalamin, determination of 423cobalt– determination of 148–151, 220–224,

232, 241–247, 251, 259, 261–267,320–322

– preservation 24, 25– sampling 13, 14continuous flow analysis– determination of anions 69, 75copper–determinationof 14, 16, 132–157, 220,

225, 229, 232, 241–248, 251–271, 316,317, 320–323

– sampling 10, 13, 14, 24, 25– speciation 157coulometry, determination of

cations 128coprecipitation techniques 353cyanide, determination of 20, 21cysteine, determination of 403

dichromate, determination of 48–53dimethyl sulfoxide, determination

of 402diuron, determination of 412dynamic headspace analysis,

determination of organics 366dysprosium, determination of

163, 196

electron spin resonance spectroscopy ofcations 157

electrostatic ion chromatography ofanions 89, 90

emission spectroscopy, determination ofanions 109

erbium, determination of 163, 196estuary waters, determination of anions

99–102ethylene diamine tetracetic acid,

determination of 401europium, determination of 163, 196

Subject Index 507

flavins, determination of 426flow injection analysis of– anions 68, 72–76, 84– cations 148, 168, 217– gases 108fluoride, determination of 53–56, 91fluorobendazole, determination of 402formate, determination of 57fractionation techniques 361free chlorine, determination of

103–108freeze drying 350, 351froth flotation 351

gadolinium, determination of 163, 196gallium, determination of 163, 262garvey sampler 4gas chromatography, determination of– anions 55, 108– cations 114, 203, 207– organic compounds 373, 377, 381,

403, 405, 411– organometallic compounds 452, 456gas chromatography-mass spectrometry,

determination of 375, 382, 411, 466gas stripping 357–361gel permeation chromatography,

determination of cations 174germanium, determination of 163Go-Flow sampler 8–11, 13, 15gold, determination of 164, 267

Harvey drum samples 4headspace sampler, determination of

gases 108–109high performance liquid

chromatography, determination of– anions 57, 58– cations 111, 209, 271, 318– organic compounds 369, 378, 384,

404, 408, 410, 461holmium, determination of 164, 196humic materials, determination of 416hydrobios sampler 9, 10, 25hydrocarbons, preservation 29hydrogen ion concentration,

determination of 21, 89, 90hydrogen sulphide, determination

of 108hydrowire sampler 8–12

hydroxy acids, determination of377–379

hydroxylamine, determination of 400hypochlorite, determination of 58

indium– determination of 164, 229, 244, 247– preservation 24, 25inductively coupled plasma atomic

emission spectrometry,determination of

– anions 86, 87– cations 139, 184, 319inductively coupled plasma mass

spectrometry, determination of– cations 123, 164, 184, 188, 214, 244,

323international council for exploration of

sea 7iodate, determination of 58–61iodide, determination of 62–65, 89, 90iodine, determination of 467ion chromatography of anions 9, 41, 64,

69, 82, 88ion exclusion chromatography,

determination of anions 455ion interaction chromatography,

determination of anions 75ion selective electrodes, determinationof– anions 47, 53–56, 109– cations 118, 138, 155, 194irgalol, determination of 412iridium, determination of 165, 230iron– determination of 165–167, 220, 229,

242–247, 260–271, 320–322– preservation 24, 25– sampling 13, 14isotachaelectrophoresis,

determination of– anions 46, 85, 86– gases 108isotope dilution analysis,

determination of– anions 72– cations 57

ketones, determination of 380, 381

lanthanum, determination of 167, 266,271

508 Subject Index

lead– determination of 15, 16, 168–174,

220–224, 229, 232, 241–258, 261, 268,269, 320–322

– preservation 75– sampling 13–15lipids, determination of 412lithium, determination of 174, 223, 268lutecium, determination of 175, 196

magnesium, determination of175–180, 229, 232, 241–247, 260–271,318, 320–323

mass spectrometry, determination ofcations 174, 192, 201, 206, 268, 319

mercury– determination of 180–186, 229–245,

253, 262, 264, 267, 317, 318– preservation 21, 25, 26– sampling 1, 14, 15metals, preservation 21microcystine, determination of 426molybdate, determination of 65, 91molybdenum, determination of 65, 91,

180–189, 229, 232, 243, 260, 262, 265

neodymium, determination of 189, 195neutron activation analysis 122, 164,

174, 180, 200, 204, 205, 206, 214, 262nickel– determination of 14, 16, 190–192,

220–224, 232, 241–247, 259–271,320–323

– sampling 9, 13, 14Niskin bottle sampler 6, 9, 10, 15nitrate, determination of 21, 65–71,

73–76, 89, 90, 99, 100, 102nitric oxide, determination of 108nitrite, determination of 21, 71, 73–5,

99, 100nitro compounds, determination of

399nitrogen compounds, determination of

393, 394nitrogen elemental, determination of

468–470non-ionic surfactants, determination of

22, 388, 389nucleic acids, determination of 397

oil spills, examination of 20, 370–377

organic compounds, preconcentration426

organoarsenic compounds,determination of 443–446

organocadmium compounds,determination of 446

organochromium compounds,determination of 49, 50

organocopper compounds,determination of 446

organolead compounds, determinationof 447

organomercury compounds,determination of 447–454

organophosphorus compounds,determination of 409–411

organosulfurcompounds,determinationof 401–403

organothallium compounds,determination of 454

organotin compounds, determination of455–461

osmium, determination of 192oxygen isotopes, determination of 192ozone, determination of 108

palladium, determination of 192pecteno toxins, determination of 423peptides, determination of 397perrhenate, determination of 76, 91pesticides, fixing 20phenols, determination of 22, 381, 382phosphate– determination of 22, 76–82, 88, 89,

100, 102– preservation 29– phosphorus, determination of 470photoactivation analysis of anions 56phthalates– determination of 382– preservation 26, 27, 29plant pigments, determination of

416–423platinum, determination of 192, 230plutonium, determination of 192, 336,

339polarography– determination of anions 42–45, 87, 88– determination of cations 151, 180,

193, 212, 318

Subject Index 509

polonium, determination of 193polyaromatic hydrocarbons– determination of 369– preservation 29polychlorobiphenyls, determination of

26, 27, 404potassium, determination of 193, 223,

262praseodymium, determination of 194,

195preconcentration of– anions 92– cations 288, 324– organic compounds 426preservation of seawater 17promethium,determinationof 194, 195propionate, determination of 82protein, determination of 397pyruvate, determination of 82

radioactive element, determination of328–348

radioactinium, determination of 328radiobarium, determination of 331radiocobalt, determination of 338radioiron, determination of 338radiolead, determination of 328, 331radiomangnese, determination of 338radiosmium, determination of 341radiophosphorus, determination of

335radiopolonium, determination of 328radioruthenium, determination of 341radiotechnetium, determination of 333radiouranium, determination of 342radium, determination of 194, 31raman spectroscopy of organic

compounds 368rare earths, determination of 194, 197reverse osmosis 350rhenium, determination of 199, 200rubidium– determination of 200, 201, 208, 223,

247, 262– preservation 24, 25ruthenium, determination of 201

samerium, determination of 195, 201sample containers 30–31sample contamination 27–33

sample preparation 349–364sampling procedure 1–17sampling devices 5–7scandium– determination of 201, 266, 267– preservation 24, 25selenate, determination of 87, 100, 101selenite, determination of 82, 83, 100,

101selenium, determination of 201–203,

229, 232–244, 260, 262, 318segmented flow analysis, determination

of anions 46sodium, determination of 204, 205solvent extraction 351–353solid state membrane electrodes, deter-

mination of anions 46speciation cations 157, 271spectrofluorimetry of– anions 82, 83, 100– cations 113, 150, 216– organic compounds 368, 370spectrophotometry of– anions 40–45, 53, 58–61, 63, 65, 73,

77–84, 88, 100, 102– cations 112, 115, 127, 148, 155, 165,

166, 186, 190, 202, 207, 211, 213, 219–organic compounds 377, 380, 381, 382– organometallic compounds 445sterols, determination of 413strontium– determination of 205, 219, 262– preservation 24, 25sulfate, determination of 56–88, 100,

102sulfide, determination of 20, 22, 85, 86,

88, 89, 102sulfur, determination of 471, 472surface films, analysis of 3–7suspended solids, determination of

90–92

technetium, determination of 205–207tellurium, determination of 205, 229,

233terbium, determination of 197, 206thallium, determination of 206, 229,

244, 269thiols, determination of 402

510 Subject Index

thorium, determination of 206, 207,265, 266, 337

thullium, determination of 196, 207tin,determinationof 208–212,229, 233,

265, 318, 319titanium, determination of 211, 262tungsten, determination of 211

ultraviolet spectroscopy, determinationof anions 66, 67, 99

uranium– determination of 211, 212, 260,

262–270, 322– preservation 24, 25urea, determination of 400

valerate, determination of 88vanadium, determination of 213, 214,

229, 243, 247,260, 267, 268vitamins, determination of 423voltammetry, determination of organic

compounds 380

volatile organic compounds,determination of 392, 393

World Meterological Commission 7

X-ray emission spectrometry,determination of anions 45

X-ray fluorescence spectroscopy,determination of

– anions 45– cations 124, 189, 201, 261– organic compounds 411

ytterbium, determination of 196yttrium, determination of 215, 233, 262

zinc– determination of 215–218, 220, 233,

242–250, 254, 261, 265–269, 320–322– preservation 22–25– sampling 13, 14zirconium, determination of 218–220,

262

AnalysisofSeawater A Guide for the Analytical and ...dl.mozh.org/up/Analysis_of_Seawate.pdf · T.R....

Documents

Transcript of AnalysisofSeawater A Guide for the Analytical and ...dl.mozh.org/up/Analysis_of_Seawate.pdf · T.R....